JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 98, NO. E9, PAGES 17,149-17,181, SEPTEMBER 25, 1993 Lunar Impact Basins: New Data for the Western Limb and Far Side (Orientale and South Pole-Aitken Basins) from the First Galileo Flyby JAMES W. HEAD 1, SCOTI MURCHIE 1, JOHN F. MUSTARD 1, CARLE M. PIETERS 1, GERHARD NEUKUM 2, ALFRED MCEWEN 3, RONALD GREELEY 4, ENGELBERT NAGEL 2, AND MICHAEL J. S. BELTON 5 Compositional aspects of impact basin materials can be analyzed using multispectral image data acquired by the Galileo solid state imaging (SSI) experiment during the December 1990 lunar encounter. These data provide important information on the spectral properties of the western lunar limb and parts of the far side. The SSI images cover the wavelength range 0.4-1.0 Bm, allowing measurement of spectral slope and estimation of the strength of the 1 Bm absorption due to iron in the mafic minerals olivine and pyroxene. Among deposits of the 930-km-diameter Orientale basin, exterior ejecta comprising the Hevelius Formation is relatively homogeneous and spectrally similar to mature Apollo 16 soils, suggesting an upper crustal source. The centrally located Maunder Formation is distinct from the younger mare basalts but comparable to the Hevelius Formation in its spectral reflectance properties, supporting an interpretation as basin impact melt. The Montes Rook Formation, located in an annulus between the Maunder and the Hevelius, shows a slightly stronger mafic absorption and may be the deepest crustal material excavated. The distal Orientale deposits show local mafic enhancements (in the Schiller-Schickard and Mendel-Rydberg regions) interpreted to represent pre-orientale mare deposits, or cryptomaria, intermixed with overlying basin ejecta. In this case, maria of sizes comparable to those presently observed were widespread in this region before the Orientale impact. Mixing-model analyses are consistent with the ballistic erosion and sedimentation model for ejecta emplacement in the distal regions beyond the continuous ejecta deposit. On the southern lunar farside, a huge area with an enhanced mafic absorption corresponds to the interior and rim of the pre-nectarian South Pole-Aitken impact basin, 2000-2500 km in diameter. The anomaly is interpreted to be due to several factors, including excavation into the more mafic lower crust, and the presence of extensive early volcanic fill (cryptomare), similar to that seen in ancient basins such as Smythii and Australe. These results show that although basin-forming events are an important factor in producing lateral heterogeneities in crustal composition, and in modifying preexisting deposits (such as cryptomaria), the majority of material in even the largest basins was excavated from crustal levels. Our results suggest a gradational vertical crustal stratigraphy consisting of an uppermost mixed crustal layer of anorthosite, basin ejecta, and cryptomaria deposits (generally corresponding to the megaregolith), an upper crustal layer of anorthosite, and a lower more noritic layer. Many of the basic questions remaining from this study could be addressed by global high-resolution geochemical and mineralogical data obtained by polar orbiting spacecraft. 1. INTRODUCTION The study of impact basins provides information on several fundamental questions in lunar science. Analysis of basin deposits yields data on processes of basin formation and evolution, including depth of excavation of basins, the nature of the emplacement of ejecta, the origin of the morphology of various ejecta deposits, the mixing of ejecta with the preexisting substrate, the origin of basin rings, and the stratigraphy of basin deposits (Figure 1). Since the formation of single impact basins has such a widespread effect on the preexisting surface (essentially complete destruction in the basin interior and ballistic modification and blanketing in the exterior), large basins can radically alter the geography of extremely large portions of the Moon (Figure 2), covering and compositionally obscuring previous craters, basins, and 1Department of Geological Sciences, Brown University, Providence, Rhode Island 2DLR, Institute for Planetary Exploration, 12484 Berlin and 82230 Oberpfaffenhofen, Germany. 3U.S. Geological Survey, Flagstaff, Arizona. 4Department of Geology, Arizona State University, Tempe. 5National Optical Astronomy Observatory, Tucson, Arizona. Copyright 1993 by the American Geophysical Union. Paper number 93JE01278. 0148-0227/93/93JE-01278505.00 17,149 volcanic deposits. For this reason, detailed study of basin deposits is required to assess the nature of prebasin geography and geology, and to reconstruct the complex early history of the Moon. In addition, impact craters and basins can be used as probes of crustal stratigraphy [Pieters, 1986]. If the geometry of the impact basin cavity and the mode of emplacement of ejecta are generally known, then the compositional characteristics of basin deposits can be interpreted in terms of vertical crustal stratigraphy in the region of impact, and lateral heterogeneities in the crust between different basins [Pieters et al., 1992]. The Galileo spacecraft encountered the Moon in December 1990 in the first of two Earth-gravity assists as part of its planned trajectory to reach Jupiter orbit in 1995 [Belton et al., 1992a]. Fortuitously, the geometry of the encounter was such that the solar illumination was complementary to that during the Apollo missions and similar to that during Zond missions [Shevchenko, 1980; Shevchenko and Chikrnachev, 1992]; the western near side, the western limb, and portions of the far side were illuminated, and the subspacecraft point passed through the interior of the Orientale basin. The Galileo solid state imaging system (SSI) was used to obtain multispectral images in the range 0.4-1.0 tm of this portion of the Moon, complementing previous Earth-based data [Spudis et al., 1984; Hawke et al., 1991] and Apollo orbital data [Metzger et al., 1974], and providing compositional information for much of the western limb and far side for the first time.
17,150 HEAD ET AL.: LUNAR IMPACT BASINS ß BASIN FORMATION AND MODIFICATION PROCESSES -Depth of Excavation -Origin of Units, Rings ß MANTLE HETEROGENEITY -Mare Location, Variability -Cryptomarla. -Thickness -Lateral Variation ß EJECTA EMPLACEMENT PROCESSES -Mixing -Nature of Substrate -Cryptomarla Detection Rim (A) Outer Edge of Radially Textured Ejecta Deposit (B) x Outer Edge of x Large Secondary x \ Crater Field (C) A)x x \ \\x (c) Fig. l. Characteristics of lunar impact basins and relation to important scientific problems. Not to scale. i I Preliminary analysis of these data reveals evidence for distinct near side-far side compositional variations, the characteristics of mare deposits on the limb and far side, the presence of ancient cryptomaria covered by younger basin deposits, and the compositional affinities and distribution of the Orientale impact basin deposits [Belton et al., 1992b]. In the present contribution, we examine the detailed nature of impact basins on the western limb and farside and address many of the questions outlined above (Figure 1). We first review the known characteristics of Orientale and other impact basins, then present spectral parameter images and a classification of units using the Galileo data to establish compositionally linked spectral units. We then compare these to morphologic units, and outline a general stratigraphic sequence of basinforming events for this region. Finally, we examine the implication of these units and the stratigraphy for the formation of impact basins, the emplacement of basin ejecta, and the stratigraphy of the lunar crust. Analyses of Galileo data relevant to crustal diversity [Pieters et al., this issue], mare deposits [Greeley et al., this issue], and lunar craters and soils [McEwen et al., this issue] are presented elsewhere. 2. THE CHARACTERISTICS OF IMPACT BASINS ON THE WESTERN LIMB AND FAR SIDE Fig. 2. Lunar hemisphere cross section with the boundaries of Orientale basin and its ejecta deposits compared to those for Imbrium and the South Pole-Aitken basin. Basin centers are at top; thus onehalf the basin is shown. Basin rim is the Cordillera ring for Orientale, the Apennine ring for Imbrium, and the 2500-km ring for South Pole- Aitken. South Pole-Aitken deposit estimates are scaled from Orientale and Imbrium. undulatory fissured terrain and is interpreted as impact melt [e.g., Head, 1974a; Head, 1977] or early high-albedo volcanic plains [e.g., Schultz, 1972]. The Montes Rook Formation (MR in Figure 6) consists of generally equant knobs 2-5 km across and associated smooth to rolling terrain, differing texturally from the generally radially textured exterior deposits of the Hevelius Formation. Most of the Montes Rook Formation Basins and large craters on the western limb and far side (Plate 1 and Figure 3) have excavated material from varying depth over a large area of the lunar crust, emplacing it on the occurs between the Cordillera and Outer Rook Mountains, surface as ejecta units, and thus exposing vertical crustal stratigraphic units in lateral ejecta deposits, so that they are visible to the Galileo imaging system. The youngest of these basins is Orientale (Figure 4; labelled O in Figure 5), and it is though in several places it extends beyond the Cordillera. The Montes Rook Formation has been interpreted as an ejecta facies derived from greater depth than ejecta exterior to the basin, and thus emplaced in a different mode and/or stage in the the largest basin whose interior deposits have largely escaped excavation process [McCauley, 1977; Scott et al., 1977; burial by mare deposits and ejecta of younger basins. Montes Cordillera (C in Figure 6), an outer basin ring with a diameter of 930 km, consists partly of massifs and partly of an inward facing scarp. Two inner massif rings with diameters of 480 and 620 km, the outer and inner Rook Mountains (OR and IR in Hodges and Wilhelrns, 1978]. This interpretation places the rim of the excavation cavity at the Cordillera Mountains. Alternatively, the Montes Rook Formation is interpreted as deposits that were initially like the radially textured Hevelius, but were modified texturally by late or postbasin megaterrace Figure 6), are distinct in the eastern part of the basin but in the formation [Head, 1974a,;ttead, 1977]. This interpretation west merge with each other and outwardly with Montes Cordillera (Figure 4) [Scott et al., 1977; Wilhelrns, 1987]. These basin rings approximately bound three morphologic units that compose the basin interior and exterior ejecta deposits [Moore et al., 1974; Head, 1974a; Scott et al., 1977; places the rim of the transient cavity at about the Outer Rook Mountains. The exterior ejecta deposits, the Hevelius Formation (Figure 4; HV in Figure 6), extend about one basin diameter beyond the Cordillera Mountains. The proximal Hevelius Formation consists of a deposit of continuous ejecta, Wilhelrns et al., 1979; Wilhelrns, 1987] (Figure 4). The textured radially to the north and south of the basin and Maunder Formation (Figure 4; MA in Figure 6), located interior to the Rook Mountains, consists of light smooth plains and exhibiting concentric ridges to the east and west. Distal portions are characterized by a discontinuous cover of smooth! / / I I I
HEAD ET AL.: LUNAR IMPACT BASINS 17,151 plains deposits and preexisting terrain heavily cratered by secondaries. Smaller, older basins and large craters located within the distal Hevelius Formation, including Schiller- Zucchius and Schickard to the southeast, contain light plains morphologically resembling mare deposits but with a highland-like albedo. These characteristics are similar to the configuration of deposits surrounding the Imbrium basin (the radially textured Fra Mauro Formation) and nearby distal deposits (Cayley Formation and secondary crater clusters) [Wilhelms, 1987; Head, 1974b, 1975; Spudis et al., 1988]. In the case of Imbrium, the distal plains have been interpreted to be non-mare volcanic plains younger than the Imbrium basin [Soderblom and Lebofsky, 1972], prebasin volcanic plains [Hartmann and Wood, 1971; Hawke and Head, 1978], Orientale primary ejecta [Chao, 1973], and Imbrium distal ejecta facies consisting largely of primary ejecta [Eggleton and Schaber, 1972] or local material mobilized and mixed by Imbrium ejecta [Oberbeck et al., 1975; Head, 1974b]. Orientale's large size (>900 km diameter) suggests the possibility that it excavated and exposed deep crust or mantle, but Earth-based telescopic observations [Spudis et al., 1984; Hawke et al., 1991] indicate a low-mafic, anorthosite to anorthositic norite crustal composition in the basin interior and ejecta that are interpreted to be derived from materials in the middle to upper crust. The only mafic-rich regions recognized prior to Galileo occur in the distal Hevelius Formation southeast of Orientale, in Schickard and Schiller- Zucchius. Dark-halo craters in this region exhibit marelike spectra [Hawke and Bell, 1981; Bell and Hawke, 1984], interpreted as being due to excavation of a pre-orientale cryptomare [Head and Wilson, 1992] partly covered by and mixed with basin ejecta [Schultz and Spudis, 1979]. Initial analyses of Galileo images of Orientale basin materials [Belton et al., 1992b; Mustard et al., 1992] confirm the relatively uniform low mafic content of most Orientale deposits, and the regionally higher mafic content of the distal Hevelius Formation within Schickard and Schiller-Zucchius. An additional region of enhanced mafic content, possibly a cryptomare, was also identified in the Mendel-Rydberg basin. Specific investigation of the interior of Orientale basin [Murchie et al., 1992] reveal subtle spectral heterogeneities of much lower magnitude than those observed within the maria. Additional basins have been mapped in the region south and west of Orientale. Of these, six (South Pole-Aitken, Apollo, Mendel-Rydberg, Bailly, Hertzsprung, and Hausen; Figure 3 and Plate 1) are well enough preserved, lack widespread mare cover, and are far enough from the terminator region in Galileo images for meaningful multicolor data covering basinforming materials to be analyzed. These basins are either smaller and less developed or older and much more degraded morphologically than Orientale, so that extensive intrabasin deposits are less distinctive or less recognizable. The largest and oldest of these is the pre-nectarian South Pole-Aitken basin (Figure 7) (SPA in Figure 5), the largest well-documented basin, which occupies a major part of the south-central far side. At least two incomplete massif rings, 2000 and 2500 km in diameter, have been recognized in Lunar Orbiter and Apollo images [Stuart-Alexander and Howard, 1970; Stuart-Alexander, 1978]. Interior to these is the largest concentration of far-side mare patches [Wilhelms, 1987]. Apollo laser altimetry showed that the basin interior is topographically lower than the surrounding highlands, and gamma ray spectroscopy demonstrates that its iron content is enriched relative to the surrounding highlands [Metzger et al., 1974; Stuart-Alexander, 1978]. Initial analysis of Galileo images demonstrated that the basin interior also has a strong 1-1 m absorption interpreted to be due to mafic minerals, possibly originating by impact excavation of lower crust or mantle material or formation of cryptomaria [Belton et al., 1992b]. The pre-nectarian Apollo basin (A in Figure 5) is superposed on the northeastern part of South Pole-Aitken. Two major basin rings have diameters of 250 and 505 km, and the interior of the basin exhibits two sizable patches of superposed maria [Hartmann and Wood, 1971; Stuart-Alexander, 1978]. Nectarian-aged Mendel-Rydberg and Bailey basins (M and B in Figure 5), located south of Orientale, are buried by the distal part of the Hevelius Formation. Mendel-Rydberg, whose morphology has been severely degraded by superposition of the Orientale ejecta, exhibits a major inner ring 200 km in diameter. Several outer rings are more poorly defined. Bailey's inner and outer rings are 150 and 300 km in diameter [Hartmann and Wood, 1971; Wilhelms et al., 1979; Wilhelms, 1987]. Hertzsprung (H in Figure 5) is Nectarian in age, and located northwest of Orientale. The major inner and outer rings are 265 and 570 km in diameter. Approximately the southeastern half of the basin interior and ejecta are buried by the distal Hevelius Formation, but the northwestern parts are exposed [Hartmann and Wood, 1971; Scott et al., 1977; Wilhelms, 1987]. Hausen (h in Figure 5) is Upper Imbrian in age [McEwen et al., this issue], 167 km in diameter, and is superposed on Bailly and the southern part of the Hevelius Formation just beyond the outer ring of South Pole-Aitken. Strictly speaking it is a crater, but it is transitional to basins in that it contains a central peak ring [Wilhelms et al., 1979; Wilhelms, 1987]. Using the Galileo multispectral images, we address two major sets of questions about these basins, focusing on Orientale due to its size and state of preservation, and South Pole-Aitken because of its extremely large size and location on the far side. First, what are the character and spatial distribution of subtle spectral variations in the Orientale basin deposits? Are these variations correlated with the morphologically defined basin materials, and if so how do the materials differ? To investigate this set of questions, we have analyzed in detail the immediate region of Orientale (Figure 8). Second, are similar heterogeneities observed in and around the other basins and basinlike craters in the southeastern farside? What are the similarities to and differences from the heterogeneities observed at Orientale? Do these observations suggest a particular geologic history of basin formation? What information do these observations provide about the stratigraphy of the lunar crust? 3. ANALYTICAL PROCEDURES Data used here are coregistered, mosaicked, geometrically reprojected SSI images, normalized to 0 ø illumination, emission, and phase angle using nonwavelength dependent Hapke parameters. A secondary correction of both color and phase darkening was applied, correcting the data to match LUNMAP 12, 19.5 ø phase [McEwen et al., this issue; Helfenstein and Veverka, 1987]. Channels whose calibration could best be verified (0.41, 0.56, 0.69, 0.76, and 0.99 I m) were analyzed spectrally. Two independent methods were applied to measure the spatial variations of spectral properties in an image format, complementing analysis of individual spectra.
17,152 HEAD ET AL.: LUNAR IMPACT BASINS Plate 1. Galileo albedo map of the highlands of the western limb and far side with the maria removed (black). Unit brightnesses are defined relative to MH0 (Mare Humorum 0) [see Pieters et al., this issue]. White (>3.10 MH0); light pink (2.45-3.10 MH0); medium pink (1.85-2.45 MH0); dark purple (1.45-1.85 MH0); black (<1.45 MH0). All mare materials fall in the lowest albedo slice, but are color-coded on the basis of UV/VIS to represent increasing TiO 2 content (red represents low TiO 2 content, through yellow, green and blue, which represents high TiO 2 content). Fig. 3a. Airbrush version of the region in Plate 1.
HEAD ET AL.: LUNAR IMPACT BASINS 17,153 //Oceanus ImbrlUm K/, / / Prøcellarum Hertzsprung Basin I Korolev / Mendel- r--,,e--schickard / Sout.o,e-,ten..,n k_dber; Basra Schiller... /..,n- Fig. 3b. Sketch map of the locations of major features in the region of Plate 1 and Figure 3a. Spectral pararneterization. In order to represent synoptically the spectral variability in the multicolor Galileo images, synthetic spectral parameters describing major sources of variance were calculated (Figure 9a). The physical origins of these spectral variations are discussed in detail by Pieters et al. [this issue]. "Visible spectral slope" is calculated as the ratio of the violet (0.41 Bm) and green (0.56 Bm) channels; "continuum spectral slope" is calculated as the ratio of the green and 0.76- tm channels. Spectral slope is a function of Ti content in mature mare soils, soil composition, and maturity. "Estimated depth of the mafic absorption band" at 1 Bm is calculated as the ratio between reflectance at 0.99 Bm and the continuum extrapolated from the green and 0.76 Bm channels. The depth of the 1-Bm absorption present in many lunar materials is affected largely by the abundance of Fe-bearing mafic minerals (pyroxenes and olivine) at the optical surface of the Moon. True measurement of the depth of the mafic absorption band is not possible in Galileo's spectral range because of the absence of a channel at the long-wavelength wing of the absorption (-1.3 Bm). However, the estimated absorption band depth, while not providing an accurate absolute measure, is a good indicator of the relative strengths of the mafic mineral absorptions. Parameter images calculated for the Orientale region are displayed with a reference shaded relief map in Figure 10. Inspection and comparison to albedo in Figure 8 reveal that major variations in continuum spectral slope in the highlands are largely related to exposures of fresh crater material, which has a less red continuum slope (brighter in Figure 10c). In contrast, variations in visible spectral slope and the estimated depth of the mafic absorption (Figures 10b and 10d) form irregular annuli around the basin, correlating at least in part with the morphologic units. The compositional differences of these morphologic units can be meaningfully inferred from spectra if two conditions are met. First, the units must be sampled by sufficient pixels to minimize any spurious differences between channels of the Galileo camera which + + + + + + + + + + _+ + + + +++++++ + +++0 Leuschner+ + + ++ + +, - + + + + +++0++ + + + + + + + + + + t-t+ + + + + + + + O + + + +/ - + -... c + + + + + + APPROXIMATE SCALE AT CENTER 0 1000 KILOMETERS I I I I I I EXPLANATION Mare materials I-'-'1 Post-Orientale crater materials.'..; +_ Orientale Group. r' Pre-Orientale terra Fig. 4. Geologic map of the Orientale Group; geologic units mapped on the basis of photographs and morphology [from McCauley, 1977]. From left to right in Orientale Group bar (Maunder Formation, double vertical bars; Montes Rook Formation, check marks are knobby facies and dots are massifs; Hevelius Formation, crosses are inner facies and dashes are outer facies; Hevelius Formation, crosses are inner facies and dashes are outer facies; transversely structured parts of the Hevelius Formation are small areas with wavy lines; some of the many secondary craters are black).
..............::,.... 17,154 HEAD ET AL.: LUNAR IMPACT BASINS Southeastern Farside a Reference Hop b 0.76-micron Albedo.;i::,/":*-'::., :... :," ' *... :.-.. '... *::.c--....;. i, '-'" ':... : :" * '".:. ---?,... ::;:a:$ :*....,:.½a } Fig. 5. (a) Shaded relief map covering the southeastern portion of the lunar far side and the southwestern limb region. Basins and basinlike craters discussed in the text and labelled here include Orientale (O), Mendel-Rydberg (M), Bailly (B), Hausen (h), Apollo (A), Hertzsprung (H), and South Pole-Aitken (SPA). (b) Photometrically corrected mosaic of Galileo images covering the same region, obtained through the 0.76- tm filter. might be due to slight misregistration, low data number (DN) levels, and/or low levels of scattered light. Second, the morphologic units should be of comparable albedo to minimize possible systematic spectral effects of differing regolith maturity caused by exposure of fresher materials on steep slopes or superposition of fresh craters (discussed below) Statistical parameterization. We also classified observed surface regions into discrete spectral units using the statistical parameterization method of Hurtrez et al. [1991]. In this method, each spectrum in an image is reduced to two dimensions, mean reflectance in all channels and standard deviation from the mean reflectance (Figure 9b). Mean reflectance in the photometrically corrected mosaics is approximately proportional to albedo; standard deviation from neutral gray is a measure of the spectral contrast and represents the departure of the spectral shape from a neutral (i.e., flat line) spectrum. Hurtrez et al. [ 1991 ] have demonstrated this method in laboratory measurements of mixtures consisting of a small number of spectrally distinct end members (olivine, pyroxene, and plagioclase). In this simple case, the data form a ternary plot with end members on the vertices and mixtures forming linear trends between them. In a planetary surface region that consists of mixtures of constant but unknown spectral reflectance end members, this technique offers three valuable capabilities. First, the vertices Fig. 6. Lunar Orbiter IV image of the interior and exterior ejecta deposits of Orientale showing the major structural features and basin deposits discussed in the text, including Montes Cordillera (C), Rook Mountains (Inner Rook, IR; Outer Rook, OR), the Hevelius Formation (HV), the Montes Rook Formation (MR), and the Maunder Formation (MA). LO-IV-181M.
Fig. 7. Moon viewed from a point above the center of the far-side South Pole-Aitken basin. Location of South Pole-Aitken basin and tings proposed by Wilhelms et al. [1979] (outer ting) and Stuart-Alexander [1978] (inner ting) are shown. Oftentale is seen on the extreme right limb. Reference Mop Orientole Region b 0,76- micron Atbedo Fig. 8. (a) Shaded relief map coveting the Orientale region. Major structural features and basin deposits discussea n the text and labelled here include Montes Cordillera (C), Rook Mountains (R), the Hevelius Formation (Hv), the Montes Rook Formation (Mr), and the Maunder Formation (Ma). (b) Photometrically corrected mosaic of Galileo images coveting the same region, obtained through the 0.76 [tm filter.
ß. 17,156 HEAD ET AL.: LUNAR IMPACT BASINS SPECTRAL PARAMETERS SPATISTICAL PARAMETERS (n channels) Reflectance relative to sun I Visible I spectral I. / / Continuum' I // spectral I r slope Wavelength, gm Est. depth of mafic band Fig. 9a. Schematic representation of the spectral parameters representing the major sources of spectral variability. Reflectance relative to sun Standard Deviation =,/( ( - r)2)/n ]!--i... 0.3 014 015 016 017 018 0'.9 1'.0 111 Wavelength,/xm Fig. 9b. Schematic representation of the statistical parameters used to classify distinct spectral units. of a mean-standard deviation diagram represent the spectral reflectance end members. Second, spatial units defined from domains on the diagram will represent the spectral equivalent of a constant range of mixtures of these end members. Third, if regolith maturation produces systematic spectral reflectance variations, which can be defined on the mean-standard deviation plot, these variations can to some extent be "seen through" to determine spatial distributions of distinct or related units. The procedure for doing this is described below. Figure 11 shows in medium gray the mean-standard deviation diagram of all spectra in the Orientale region. We selected homogeneous northern and eastern portions of the Hevelius Formation as being representative of basin-related deposits, and identified within these areas several fresh, small craters which were assumed to be less mature regolith formed on the same lithology. For each crater, 50-100 pixels with a wide range of brightnesses were measured. As typified by the plot in white in Figure 11, increasing mean reflectance is accompanied by only a small systematic increase in standard deviation. If spectral shape scaled linearly in proportion to increasing brightness, standard deviation would increase proportionally to mean reflectance. The reason for this behavior is apparent in a Reference k4op b Visible Spectrot Slope ':-i '* ' ":'.:,. :- i' --::½? : ;: ½ ;< :;:,:: :-.... : -.....:% ;;: ;......:,..&'*: "i½'; : %...:..,..: C Continuum Spectrol Slope d Est, Depth of ofic Jon- ' Fig. 10. Spectral parameter images of the Orientale region. (a) Shaded relief map for reference. Major structural features and basin deposits discussed in the text and labelled here include Montes Cordillera (C), Rook Mountains (R), the Hevelius Formation (Hv), the Montes Rook Formation (Mr), and the Maunder Formation (Ma). (b) Visible spectral slope, calculated as the ratio of reflectances in the violet and green channels. Brighter tones correspond to less red spectral slopes. (c) Continuum spectral slope, calculated as the ratio of reflectances in the green and 0.76-gm channels. Brighter tones correspond to less red spectral slopes. (d) Estimated depth of the mafic band, calculated as the difference between reflectance at 0.99 gm and the continuum extrapolated from the green and 0.76-gm channels. Brighter tones correspond to stronger absorptions.
HEAD ET AL.' LUNAR IMPACT BASINS 17,157 Oreqigto Req on (Trend Due to 'Regotil, h Maturedion') (Trend Due o Exposure of Fresh Marie Material) Bean Refleclance Fig. 11. Plot of mean reflectances and standard deviations for all pixels covering the Orientale region, in medium gray. The trend shown in white represents difference in rcgolith maturation on a single lithology. The trend shown in light gray represents mixing of bright marie material excavated by a fresh crater with marie-poor material comprising the Hcvclius Formation. Figure 10. In general, as albedo of fresh craters increases, continuum slope becomes much much less red, visible slope becomes slightly less red, and the estimated depth of the marie absorption increases slightly. These broad changes are discussed in detail by McEwen et al. [this issue]; they all have the effect of making a spectrum flatter, and supressing the increase of standard deviation with albedo. In contrast, a different behavior can occur where distinct materials mix together. An example of this is shown by the points in light gray in Figure 11 which represent a fresh crater near the Schickard cryptomare that exhibits a strong marie band presumably due to excavation of mafic-rich material. The standard deviation for this crater actually decreases with increasing mean reflectance, forming a trend distinct and nonparallel to that resulting from regolith maturation. The decreasing standard deviation results from the large increase in marie absorption strength due to intermixture of marelike material. This exercise shows that regolith maturation on representative highland surfaces can be represented as a linear trend on a mean-standard deviation diagram. Mixing of optically distinct materials can form a distinct and nonparallel trend. If regoliths on the different basin deposits mature similarly, then units classified by slicing the mean-standard deviation diagram parallel to the observed regolith maturation trend should be related more to properties of the substrate, minimizing differences resulting from regolith maturation. Following these principles, we classified the Orientale region into spectral reflectance units. The units were defined as domains on the mean-standard deviation diagram, derived by slicing the plot primarily parallel to the trend due to regolith maturation. Specific unit boundaries on the diagram were derived by iterating the slicing procedure, until mapping of these domains onto the image mosaics produced the most spatially coherent units. Secondary slicing of the diagram on the basis of albedo was performed to isolate spectral reflectance end member regions. The procedure was iterated until we identified a small number of spatially coherent units whose boundaries are verifiable independently on the basis of one or more parameter images in Figure 10. Nine units were identified, three representing different regions with stronger mafic bands and six representing highland regions with weaker mafic bands (Plate 2). Spatially, these spectral units correspond in large degree to underlying morphologic units (Plate 3; compare to Figures 4 and 6). Of the more mafic units, "low-albedo mafic material" correlates in this region with mare deposits, "higher-albedo marie material" with light plains interpreted as cryptomarc, and "fresh mafic craters" with fresh craters in the maria and cryptomaria. The six less mafic units correspond approximately to higher-albedo parts of the Inner Rook Mountains, the Maunder Formation plus lower-albedo parts of the Inner Rook Mountains, the Montes Rook Formation, proximal parts of the Hevelius Formation, distal parts of the Hevelius Formation, and fresh craters with moderate to weak mafic bands. Figure 12 shows spectra of these materials, normalized to the unit corresponding to the distal Hevelius Formation to accentuate subtle variability. Average normalized spectra of the less mafic units are shown in Figure 12a, and a selection of higher-albedo mafic regions including cryptomaria and an example of a fresh mafic crater are shown in Figure 12b. The success of using this technique to incorporate differences in regolith maturity may be judged by comparing the spatial distribution of spectral units in Plate 3 with albedo variations due to fresh craters in Figure 8. Most fresh craters have little effect on the classification of surrounding regions superposed by rays or bright ejecta. The major exception is the very bright crater Byrgius (in the east-central part of the region), which differs from other fresh craters in being unusually blue and mafic. Despite some success in determination of spatial heterogeneities in substrate spectral properties, however, this technique does not remove effects of regolith maturation from the spectra themselves. For this reason, comparative compositions of fresh and mature soils remain difficult to infer. Comparison of different basins. These two analytical procedures were next applied to the southeastern parts of the far side, using the same parameterizations as for the smaller Orientale region, to produce a directly comparable set of observations of different basins. Extrapolating the spectral classification derived for Orientale to new regions is valid if the end members being mixed are the same, so parameterized spectral properties of the two regions were compared to corroborate the unit mapping. Spectral parameter images for this area are shown in Figure 13. The unit map derived from the mean-standard deviation diagram is shown in Plate 4, together with major physical geologic features including basin rings and outer contacts of basin ejecta. This representation allows direct comparison of units with underlying morphology. Average spectra of patches classifying as distinct units are shown in Figure 14, in comparison with spectra of analogous units at Orientale. Spectral mixture modeling. In some cases, we wished to detect the presence of cryptomaria (mixtures of prebasin mare material and highlands basin ejecta), and to determine the
17,158 HEAD ET AL.: LUNAR IMPACT BASINS Orientale Region MORœ MAFIC Lo -olbedo laofic Higher-albedo blofic Fresh Uofic Crater LESS MAFIC Bright Inner Rook Idts. Maunder Fm,/Rook Mrs. Montes Rook Fm, Pro%imoI Hevelius Fm, D;stol Hovel;us Fm, Mean Reflectance Fresh Croter Plate 2. Spectral units in the Orientale region. Units were classified by slicing the diagram of mean reflectance and standard deviations primarily parallel to the trend representing regolith maturation on a single lithology, and secondarily on the basis of albedo. The legend shows the geographic correspondences of the spectral units. UNITS SYMBOLS MORE MAFIC Low-albedo Marie Higher-olbedo Mofic Fresh Mofic Croler LESS MAFIC Bright Inner Rook Maunder Frn,/Rook Montes Rook Fm. Proximol Hevellus Frn. Distal Hevellus Fm Fresh Croier.-Basin Eiecto Contact... Exposed Bosin Ring... Buried Bosin Ring Plate 3. Spectral unit map of the Oftentale region. Units correspond to those depicted in the mean-standard deviation diagram in Plate 2. The legend shows the geographic correspondences of the spectral units Major basin tings and basin ejecta contacts are compiled from Scott et al. [1977], Stuart-Alexander [1978], and Wilhelrns [1987], and are shown in Figures 4 and 6.
Orientale Basin Materials Circum--Orientale Mafic Patches (Relative to Distal Hevelius Fl.) Proximal Hevelius - Maunder Fl. Montes Rook Fl. (Relative to Distal Hevelius Fl.) - Hevelius Fl. (ref.) Mendel-Flydberg IL plains O C C C O - Distal Hevelius Schickard It plains Inner Rook Mtns. SW pre--nectarian terra Fresh Craters Unnamed crater (ref.) (.a). 0.30 0:50 0.70 0.0 1.10 Wavelength In Microns (b),, I,, I., I m 0.30 0.50 0.70 0.90 1.1 Wavelength In Microns Lacroix C Fig. 12. Galileo spectra of deposits in the Orientale region, scaled to the integrated spectrum of distal parts of the Hevelius Formation to enhance subtle spectral variability. Spectra are scaled to unity in the green channel and offset vertically. (a) Integrated spectral properties of the less marie spectral units depicted in Plates 2 and 3. (b) Spectra of specific higher-albedo marie materials, and an example fresh mafic crater located in the Schickard cryptomare. Spectra of the proximal Hevelius Formation and a small fresh crater within it are shown for reference. Proximal Hevelius spectrum is shown so that the cryptomare spectra can be compared with Oftentale ejecta that are relatively uncontaminated by any mafic-rich components. b Visible Spectra( Slope.. '"? :.i...,"'"' '... -...: :.':F:.. %.:.-. --::.. :...':'.....::..... :.. :. :........ ::..:...:-..:..:...:..::' ::.. :' ' ß.,.": ;:::.:.f. ::.' C Continuum Spectral Slope d Est. Depth of Mofic Bond Fig. 13. Spectral parameter images of the Orientale region. (a) Shaded relief map for reference. Basins and basinlike craters discussed in the text and labelled here include Orientale (O), Mendel-Rydberg (M), Bailly (B), Hausen (h), Apollo (A), Hertzsprung (H), and South Pole-Aitken (SPA). (b) Visible spectral slope, calculated as the ratio of reflectances in the violet and green channels. Brighter tones correspond to less red spectral slopes. (c) Continuum spectral slope, calculated as the ratio of reflectances in the green and 0.76-gm channels. Brighter tones correspond to less red spectral slopes. (d) Estimated depth of the mafic band, calculated as the difference between reflectance at 0.99 gm and the continuum extrapolated from the green and 0.76-gm channels. Brighter tones correspond to stronger absorptions.
.. 17,160 HEAD ET AL.: LUNAR IMPACT BASINS UNITS... MORE MAFIC Low-albedo Mof;c Higher-albedo Marie Fresh Mof-Jc Crater LESS MAFtC -'SYMBOLS Brlght Inner Rook Mrs Maunder Fl./Rook Mrs Monies Rook' Fro, Proximal HevelJus Fl. Distal Hevel ul) Fl. Fresh Croter ß.i------- Bos;n Ejecto Contact... E posed Bosin Ring. '., -... Burled Bo' Jn R;ng `.*.m,... ',, ø.... ø:'",t, -,..,..,. -SPA Basin' Ring Plate 4. Spectral unit map of the southeastern far side. Units correspond to those depicted in the mean-standardeviation diagram in Plate 2, and the spectral unit map of the Orientale region in Plate 3. Basin rings and basin ejecta contacts are compiled from Scott et al. [1977], Stuart-Alexander [1978], and Wilhelms [1987]. Only the outermost well-defined rings are shown, except in the case of Mendel-Rydberg where the only well-defined ring is depicted. Comparison of Basin Materials --r- -'-r--'r--t---r- I ' i I ' 5 ; (Relative to Distal Hevelius Fl.).),, Total Hevelius Formation Hertzsprung Ejecta Montes Rook Formation Hertzsprung Interior Hausen Hertzsprung Inner Ring Inner Rook Mountains accurate applications of nonlinear models require that the data be calibrated to a very high degree of photometric accuracy. In a linear mixture model, the endmember abundances are insensitive to calibration, as long as the data are related to true reflectance by a linear gain and offset correction. Given a data set calibrated to absolute reflectance, the primary difference between the results of a nonlinear versus a linear mixture approach is the absolute magnitude of the endmember abundance values. The general spatial associations and systematics will be similar. 4. SPECTRAL PROPERTIES OF ORIENTALE BASIN MATERIALS i I I I J I 0.30 o.so o.7o o.o Wavelength In Miorons Hertzsprung Central Patch Maunder Formation Fig. 14. Galileo spectra of basin units on the southeastern far side, scaled to the integrated spectrum of distal parts of the Hevelius Formation to enhance subtle spectral variability. The spectra from Hertzsprung are shown in comparison to the corresponding units at Orientale. amount of mare material that might have been locally excavated and admixed in the ejecta emplacement process. For this purpose, an image-based, linear spectral mixture model [Singer and McCord, 1979; Adams et al., 1986] was used to determine the spatial distribution and abundance of surface spectral components in the Galileo SSI data. Since the components are anticipated to be intimately mixed on the surface, a nonlinear model may be more appropriate [e.g. Pieters et al., 1985; Mustard and Pieters, 1989]. However, Results from spectral parameterization. Materials comprising the interior and ejecta of Orientale exhibit much less heterogeneity than do other regions, such as the maria (Figure 13). However, there are subtle differences between major morphologic units, and these units are large enough and similar enough in albedo for meaningful comparison. The Maunder Formation and low-albedo parts of the Inner Rook Mountains are characterized by a very red visible spectral slope (dark in Figure 10b) and very weak mafic absorption (dark in Figure 10d). The proximal part of the Hevelius Formation closely resembles the Maunder Formation, but has an overall redder continuum slope. A partial annulus approximately correlating with the Montes Rook Formation has a slightly stronger mafic absorption and a less red visible spectral slope than either of these units. Parts of the Inner Rook Mountains have a much higher albedo, making detailed comparison with the previously discussed units difficult; their main spectral difference is an overall less red spectral slope. The Hevelius Formation exhibits both regionally subtle and locally strong spectral heterogeneities. Distal, southern,
HEAD ET AL.: LUNAR IMPACT BASINS 17,161 and western parts of the Hevelius Formation differ from from its proximal parts in having an overall less red spectral slope and a slightly stronger mafic absorption. Three local regions with a significantly enhanced mafic absorption have been identified. The first is located in and around Schickard, in regions interpreted as cryptomaria. The second is located in the central southern edge of the region covered in Figure 10, in the northern part of Mendel-Rydberg basin. Both regions are distinguished from the surrounding Hevelius Formation by their stronger estimated mafic band depth and by the presence of small, fresh craters with anomalously strong mafic absorptions. Their spectral slopes also differ from the surrounding Hevelius: the Schickard region is redder, and the Mendel-Rydberg region is less red. Morphologically, both regions also have the common attribute of local concentrations of superposed mare patches. These new results support a cryptomare interpretation for both regions [e.g., Hawke and Bell, 1981; Belton et al., 1992b]: Orientale ejecta covered and intermixed with preexisting mare deposits, and locally were buried by subsequent volcanism. Small fresh craters have excavated concentrations of the buried mare materials. In the of spectral parameters. The orange unit in Plate 3 corresponds to parts of the Maunder Formation and lower-albedo parts of the Inner Rook Mountains. It is characterized by a relatively red visible spectral slope and weak mafic absorption (Figure 12a). The red unit, characterized by a similarly weak mafic absorption, corresponds to portions of the proximal Hevelius Formation south and especially east of the basin. It differs mainly in having a slightly lower albedo and redder continuum slope. The much higher-albedo yellow unit covers parts of the Inner Rook Mountains, and differs mainly in having an overall less red spectral slope. On the mean-standardeviation plot, these three units are separated from each other in a direction consistent with regolith maturation. Thus, their differences may only involve differing histories of regolith processes rather than lithologic differences. The light blue unit in Plate 3 corresponds in most places with the Montes Rook Formation. It is distinguished by a slightly stronger mafic absorption than in the comparably bright interior deposits and the proximal Hevelius. The distal Hevelius Formation, shown in pink, has a slightly stronger mafic absorption than both the basin interior materials and the proximal Hevelius. Spatial extents of the "high albedo mafic regions" (cyan in Plate 3) are more evident in the spectral unit map than in the band depth image in Figure 10d. The mafic-rich region at Schickard (at the southeastern corner of the map) extends approximately 200 km northwest of the crater rim, where there is a local concentration of craters resembling fresh mafic craters in the maria. This result suggests that the cryptomare deposits locally extend outside the buried basins and into the surrounding terra. The spectra in Figure 12 reveal a close similarity of this cryptomare to the interior of Mendel- Rydberg. However, the "high-albedo mafic material" in the Nectarian terra at the southwest corner of the map (in the vicinity of the South Pole-Aitken basin) is distinctive in having a weaker mafic band and redder visible spectral slope; we return to a detailed discussion of this region later. 5. DISCUSSION AND INTERPRETATION OF ORIENTALE BASIN On the basis of these data we now turn to the assessment of a series of problems linked to the formation and evolution of the Orientale basin and lunar basins in general (Figure 1). case of the Schickard region, the superposed and presumably the buried maria are relatively red; in the case of Mendel- Rydberg they are relatively blue. We will return to a more The Maunder Formation detailed analysis of this region. The third region with an enhanced mafic band depth is located at the southwestern corner This unit, located in the central portion of the Orientale of the Orientale region, at the outer limit of the Hevelius basin and overlapped and embayed by the low-albedo maria Formation. This more mafic surface was mapped by Scott et al. [Greeley et al., this issue], has been interpreted as impact melt [1977] as Nectarian highlands not buried by Orientale ejecta, deposits associated with the Orientale basin-forming event and will be discussed in greater detail below. [Head, 1974a; McCauley, 1977], or as early low-albedo lava Results from spectral unit classification. The spatial flows postdating the Orientale event, but predating the maria distributions of spectral units classified using mean and [Schultz, 1972]. Two types of data are relevant to this problem: standard deviation (Plate 3) approximately correspond with (1) compositional affinities of the unit (e.g., is there an morphologic units and with parameterized spectral variations enhanced mafic absorption that would suggest a basaltic illustrated in Figure 10. These independently derived results composition?, what is the relation of the composition to that corroborate the conclusions that there are very subtle, yet of Orientale ejecta units?), and (2) age (is the unit spatially coherent spectral variations in Orientale basin contemporaneous with the ejecta emplacement, or is it later?). materials, and that they are related at least in part to Results from both the spectral parameterization and spectral morphologic units. unit classification studies (Figure 10 and Plate 3) show that the Among the Orientale basin deposits, the major distinction compositional characteristics of the Maunder Formation appears to be that materials comprising the basin interior and (including a weak mafic absorption) are distinct from the mafic proximal ejecta are less mafic than the distal ejecta. Secondary nature of the overlying mare basalts, and indeed are distinct distinctions may be made among the interior deposits; these from the higher-albedo mafic units that characterize some of reflect the same distinctions mentioned above in the discussion the cryptomaria regions. In addition, the Maunder Formation is essentially spectrally indistinct from the proximal Hevelius Formation, the primary ejecta deposit of the Orientale basin. Near-infrared spectra from the Maunder are similar to those of mature highlands comparable to Apollo 16 soils [Spudis et al., 1984; Hawke et al., 1991]. Together with its unique surface texture and its distribution, these compositional characteristics are consistent with the Maunder Formation being comprised of impact melt generally representing the average composition of target material [e.g., Grieve et al., 1991; Grieve, 1991] in the Orientale region. On the basis of these considerations, we conclude that the Galileo data support earlier interpretations that the Maunder Formation represents impact melt emplaced during the terminal modification phase of the Orientale impact event, and that there is at present no evidence to suggest mafic mineral contents comparable to those of extrusive basalts. In order to assess the relative ages of the Maunder Formation, Orientale ejecta, and the maria, crater counts of portions of the unit were undertaken (Figure 15). These data show that portions of the Maunder Formation are comparable
- -. -- -- -.. 17,162 HEAD ET AL.' LUNAR IMPACT BASINS in age to the Hevelius Formation (about 3.8 Ga), but that other portions appear to be slightly younger. The spectral data discussed above strongly argue that the possibly younger ages do not represent high-albedo basalt flows. Other possible explanations for the crater observations include (1) the rough terrain of the Maunder Formation may make accurate counting of small craters difficult, giving a relatively younger age; (2) since impact melt is very mobile and is the last material to come to rest at the end of the short-term modification stage of the basin event, there may be a difference in crater density resulting from this phenomenon; and (3) once impact melt has been emplaced, the basin undergoes longer-term modification processes, including cooling and internal thermal contraction and subsidence of the melt sheet, as well as thermal contraction of the substrate [Bratt et al., 1985]. Some of the smooth facies of the impact melt deposits could be extruded from the cooling core of the melt sheet during this time, leading to the emplacement of melt units of slightly younger age. The Hevelius Formation Fig. 15a 50 km This unit comprises the exterior ejecta deposit of the Orientale basin [Scott et al., 1977] extending from the Cordillera Mountains outward. We first discuss its compositional characteristics, and then assess the distribution of the deposits and their relation to spectral units. Apollo orbital geochemical data exist for the northern quadrant of the basin exterior deposits and gamma-ray data permit the assessment of concentrations of Fe, Ti, Mg [Bielefeld et al., 10 0 'E 10-1 Z 9 = 10. 2 10 0 'E. 10-1 Z 9 = 10. 2 ['"'l ' ]'"'"l ' ' ',,,\Orientale Event - - / 9 0. 3 10-3, - -4 m 10 10-4 E E 10-5,,,,,,,I 10'2 10 '1 10 0 10 1 10 2 10'2 10-1 10 o 10 1 10 2 Crater Diameter D(Km) Crater Diameter D(Km) Fig. 15b Fig. 15c Fig. 15. Impact crater size-frequency distribution data for the Maunder Formation and adjacent Mare Orientale. (a) Southcentral Orientale basin interior (Lunar Orbiter IV 195 H1) showing location of counting areas. Maria areas in this image have crater model ages of 3.45-3.70 Ga [Greeley et al., this issue], while the Orientale Basin ejecta is about 3.8 Ga. (b) Crater sizefrequency distribution data for bright smooth plains facies of the Maunder Formation (area 1 in Figure 15a) Crater model age is 3.75 Ga. (c) Crater size-frequency distribution data for corrugated facies of the Maunder Formation (area 2 in Figure 15a). Crater model ages are 3.60 Ga and 3.81 Ga.
HEAD ET AL.: LUNAR IMPACT BASINS 17,163 1976; Davis, 1980], and Th [Metzger et al., 1977]. Mixing models linking these data to the chemistry of lunar rock types for the region on the northern part of the ejecta deposit [Spudis et al., 1984] showed that (1) the region could be modeled as primarily a mixture of "anorthositic" components (anorthosite, anorthositic gabbro, and gabbroic anorthosite); (2) minor amounts (<5%) of low-k Fra Mauro basalt are present; (3) minor amounts of high-ti mare basalt components are locally present; (4) there is no evidence for appreciable amounts of norite, ultramafic rocks, or high-k KREEP; and (5) there is no evidence of distinctive radial or concentric variations in the composition of Orientale ejecta. Visible and heal - tt are,, t... a of the eastern quadrant o ' the Hevelius Formation visible from Earth show that the surface is dominated by noritic anorthosites, and that the Hevelius in this region is areally uniform and very similar to Apollo 16 soils [Hawke et al., 1991 ]. Galileo data for the Orientale region as a whole are consistent with the general view that the Hevelius Formation is comparable to Apollo 16 mature soils in its general characteristics (Figures 12-14 and Plates 3 and 4). However, the data also show that there are heterogeneities at both regional and local scales. Correspondence of spectral and morphologic units. The Hevelius Formation and other basin-related ejecta subunits defined on the basis of morphology do not correspond specifically in terms of their boundaries with units defined on the basis of their spectral characteristics (compare Figure 4 and Plates 3 and 4). This is not surprising, since it is known that variations in ejecta morphology are related primarily to the mode of emplacement of basin ejecta [Oberbeck et al., 1975], rather than to its composition or depth of origin. In addition, much of the ejecta (particularly the distal portion) is thought to be derived from the upper part of the crust in the target area [Spudis et al., 1984; Hawke et al., 1992] which should be very similar compositionally to the surrounding crust on which it is emplaced. Influence of substrate. Emplacement of ejecta from craters and basins is a dynamic process involving the impact of ballistic ejecta and the excavation of local pre-basin substrate material [Oberbeck, 1975]. Thus, ejecta deposits can be used to detect the characteristics of the prebasin substrate. Spectral units with characteristics indicating a more mafic composition than typical of the proximal Hevelius Formation are observed in distal regions of the Orientale ejecta deposits. In particular, high-albedo mafic-rich plains and fresh mafic craters have been used to indicate the presence of underlying cryptomaria to the southeast (Schickard; Schiller-Zucchius) and south (Mendel- Rydberg). In addition, a border of high-albedo mafic material several tens of kilometers wide along the western edge of Oceanus Procellarum (Plate 4) suggests a zone of mixing between Procellarum and the adjacent highlands, or possibly the presence of a Procellarum cryptomare. Furthermore, there is some evidence for small cryptomare patches in the vicinity of Criiger (Plate 4), an observation that is supported by other data [Hawke et al., 1991]. Patchy distribution of high-albedo mafic units northwest of the basin are likely to be related to the distinctive mafic units southwest of Orientale in, and associated with, the South Pole-Aitken basin. Heterogeneities and compositional implications. A broadscale annulus of less-mafic units surrounds the Orientale basin and generally corresponds to the Hevelius Formation (Plate 4). The proximal Hevelius Formation unit (weak mafic band and red continuum slope) is best developed in the eastern quadrant, and interspersed with the more mafic components of cryptomaria, the southern quadrant. The northern quadrant is dominated by the distal Hevelius Formation unit (slightly greater mafic band depth), while the western quadrant is dominated by the distal Hevelius Formation unit and high albedo mafic units associated with the South Pole-Aitken basin. These trends in the eastern part of the basin are consistent with the,visible-near infrared spectra [Hawke et al., 1991] which were interpreted to indicate the presence of anorthositic materials in the Inner Rook Mountains and noritic anorthosites-anorthositic norites in the eastern, proximal Hevelius. Montes Rook Formation This unit is located primarily in an annulus between the Cordillera scarp and the outer Rook Mountains, and between the Maunder Formation (interpreted to be impact melt) and the Hevelius Formation (interpreted to be basin ejecta). The Montes Rook Formation is characterized by a distinctive knobby or domical morphology [Head, 1974a; Scott et al., 1977]. The spectral unit identified with the Montes Rook Formation (Plates 3 and 4) is characterized by a slightly stronger mafic absorption than the Maunder or the proximal Hevelius formations. The position of this inner annulus unit is consistent with it lying on the rim of the excavation cavity, if that coincides with the outer Rook Mountain ring [Head, 1974a; Spudis et al., 1984]. The unit would then represent somewhat deeper ejecta than the proximal Hevelius. The upper lunar crust has been characterized as "anorthositic gabbro" (A120 3 26-28 wt%) while the lower crust is thoughto be "noritic" (A120 3 ~20 wt%)[spudis and Davis, 1986]. This slightly more mafic material could correspond to material excavated from the uppermost portion of the lower crust. The innermost unit, the Maunder, is similar to the proximal Hevelius because it is thought to represent the homogenized average of the target material melted by the impact [e.g., Grieve et al., 1991]. In addition, because of the generally spherical cap geometry of the excavation cavity, most of the impact melt volume will be dominated by upper crustal materials. Nature and Origin of Rings For the most part, the rings of the Orientale basin are generally below the effective resolution of the Galileo multispectral data. However, portions of the inner Rook Mountains show up quite distinctly as a low-mafic unit comparable to the proximal Hevelius but with a less red spectral slope (Plates 3 and 4). This difference may simply be due to the relatively fresh nature of the materials preferentially exposed on the rugged peaklike inner Rook Mountains. When the wavelength range represented by the Galileo data is supplemented by Earth-based near-infrared data [Spudis et al., 1984; Hawke et al., 1991], it is seen that the entire eastern inner Rook Mountains contain only minute amounts of low-ca pyroxene and thus are interpreted to be composed of essentially pure anorthosite, in contrast to the compositions of most other Orientale units which are slightly more mafic. Indeed, the absence of the plagioclase absorption at 1.25 gm in some of the spectra suggests that the peaks have undergone variable, and often very high levels of shock associated with the impact
17,164 HEAD ET AL.: LUNAR IMPACT BASINS [Spudis et al., 1984; Hawke et al., 1991]. On the basis of the nature of impact cratering processes [e.g., Grieve, 1991], the distinctive spectral character and apparently highly shocked nature of the inner Rook Mountains strongly suggests that this ring originated within the cavity of excavation, and not on the crater rim or outer slump terrace. Depth of Excavation and Crustal Stratigraphy If the structure nd stratigraphy of the target material are well known (or fixed, as in the case of laboratory experiments), then they can be used to study the nature of the cratering process. Alternatively, information about the nature of the cratering process (depth of excavation, distribution and emplacement of ejecta, etc.) can be applied to an analysis of the nature of the preimpact target stratigraphy. Unfortunately, neither the basin-forming process nor the target stratigraphy are known sufficiently well for either to serve as a fixed parameter. Thus, assumptions must be made and apparent solutions iterated, and reiterated. Analysis of returned lunar samples and their spatial distribution, mineralogic and geochemical data for the lunar surface, together with models of the formation and evolution of the lunar crust, provide some scenarios for the vertical stratigraphy of the lunar crust and uppermost mantle. We adopt the chemical and petrological model of the lunar crust of Spudis rind Davis [1986], which is based on lunar samples, orbital gi:ochemical data for about 20% of the surface of the planet, and irlfercnces from the impact cratering process [Davis and Spudis, 1987]. In this model, the upper highlands crust has the bulk composition of anorthositic gabbro (A1203 26-28 wt%), while the lower portion is interpreted to be noritic (A1203-20 wt%) (Figure 16). The uppermost layer (megaregolith) consists of a mixture of impact ejecta, anorthosite, cryptomaria, and other mafic intrusive and extrusive deposits. The boundaries of all three units are gradational. In an analysis of the Orientale basin, Spudis et al. [ 1984] concluded that the excavation crater diameter was about 500-600 km with a depth of excavation consistent with a proportional cavity growth model, and that the impact excavated predominantly upper crustal rocks with the effective depth of excavation being less than about onehalf the local crustal thickness (<-50 km). Although the basin- forming impact may have excavated a very deep transient cavity, most of the deeper material is thought to be rotated and displaced within the cavity, while the excavation geometry indicates that over 90% of the ejecta volume is derived from depths shallower than 35-40 km [Spudis et al., 1984]. They also note that the composition of the Orientale ejecta is in marked contrast to that of most other lunar basins which have mixtures of anorthositic material and low-k Fra Mauro (LKFM) (upper and lower crustal material), the proportions depending on basin size and pre-impact crustal thickness. Only small basins on the far side (e.g., Milne, Korolev) where the crust is thick, display ejecta deposits that are compositionally similar to Orientale [Spudis et al., 1984]. Our data for the basin deposits as a whole (Plates 3 and 4) support these conclusions. The slight mafic enhancement in the Montes Rook Formation (interpreted to represent the deepest ejecta) revealed by the Galileo data (Plate 3 and Figures 12 and 13) suggests that Orientale may have just barely sampled lower crustal material in its external ejecta (Figure 16). The often highly shocked anorthositic peaks associated with the inner Rook Mountains [Spudis et al., 1984; Hawke et al., 1991] and found in the inner rings of several other basins [Hawke et al., 1992] are interpreted to be portions of the crater cavity located at the lowermost portion of the upper crust, least contaminated by mafic extrusives and intrusives, that were rotated inward and upward in the terminal stages of the event. The Hevelius Formation is interpreted to have been derived from the upper crust, with the purest anorthosite excavated from depths similar to the Inner Rook Mountains, and having been emplaced closer to the rim than the slightly more mafic, more shallowly derived, upper crust. Model of Basin Formation and Evolution A variety of models for the formation and evolution of impact basins, with emphasis on Orientale as a type example, have been proposed and are summarized in Wilhelrns [1987]. Although our data are not sufficient to conclusively prove a specific model (for example, they do not directly address points 1, 4, and 10), we find that they are most consistent with the following scenario (Figure 17): 1. Crustal thickness. The projectile that produced the Bulk Compositions - / Former shallow layered pluton Whole crust rocks Melt Crustal Layers pluton KREEP-basalt and flows Mg ' anorthosltes ' Mare Gabbroic anorthosite anorth. gabbro anorth. norite Vertical scale.}- --,0 Megaregolith ß FAN. -.L -40 ß Nomenclature AI O. Mantle 1. "Anorthositic riorite" 24-25% ol/px > pig 2. "VHA basalt" 20-25% 3. "LKFM basalt" 18-22% 4. "Anorthositic gabbro" 26-28% 5. "Norite" 20% Anorth. norite riorite KREEP-norite IEEP substratum 80 km Fig. 16. Model for the structure and composition of the lunar crust [after Spudis and Davis, 1986]. The contact between the "anorthositic gabbro" upper crustal layer and the "norite" lower crustal layer is gradational on the scale of kilometers.
HEAD ET AL.: LUNAR IMPACT BASINS 17,165 Cordillera Ring Outer Rook Ring ORIENTALE /,. Inner Rook Ring BASIN Hevellus Formation Montes Rook Formation I Maunder Formation ] 1 t km '...,, " 7... --_ '" "' ' Upper rust LoMwege r r r uu I Fig. 17. Interpretive cross-sectional model for the formation and evolution of the Orientale basin. The transient crater penetrates to lower crustal depths but the total volume excavated from these depths is small and is deposited on the initial crater rim. Impact melt is generated during the event (see Figure 18) through melting and mixing of the crustal column. In the latter stages of the event, the transient crater collapses along an outer ring fault (the Cordillera ring) and rim material rotates inwards and upwards, forming the Outer Rook ring (the deformed transient crater rim) and the Inner Rook peak ring. Rim ejecta (ejecta derived from the deepest depths) is deformed into the texture of the Montes Rook Formation during rim collapse, creating a textural contrast with the more distal radially textured Hevelius Formation. The scarps of the Cordillera and Outer Rook rings are composed primarily of basin ejecta underlain by megaregolith, while the scarps and peaks of the Inner Rook Mountains are characterized by deeper upper crustal material rotated from the edge of the transient crater into the basin floor by cavity collapse and listfie faulting. Impact melt (the Maunder Formation) comes to rest in the basin interior and cools, later to be flooded by mare basalts. No vertical exaggeration in basin geometry and crustal thickness, although the thickness of nearsurface units is exaggerated so that they are visible. Orientale basin impacted into a crust estimated to be about 70-100 km in thickness [Bills and Ferrari, 1976; Bratt et al., 19851. 2. General crustal stratigraphy. The crust was characterized by two broad layers with gradational contacts, the uppermost of which has the bulk composition of anorthositic gabbro (A1203 26-28 wt%), while the lower portion is interpreted to be noritic (A1203 ~20 wt%) (Figure 16) [Spudis and Davis, 1986; Ryder and Wood, 1977; Warren, 1990]. These layers are interpreted to be derived from an initial ferroan anorthosite crust produced by plagioclase flotation and olivine/pyroxene sinking in a magma ocean, with the residual liquid forming a noritic lower crust and a KREEP substratum. This initial largescale stratigraphy was complicated by subsequent intrusions (forming Mg-suite intrusions primarily in the middle to lower crust) and extrusions (basalts). 3. Uppermost crustal stratigraphy. The uppermost crustal layer contains additional subdivisions, including (1) mare basalt layers in the upper 0-4 km (depending on the geometry of the topographic regions that they are infilling [Head, 1982]), which in some cases became Orientale cryptomaria; (2) a megaregolith, which represents portions of the upper crustal unit that have been physically modified to allochthonous and autochthonous breccias to depths of tens of kilometers by crater and basin impact reworking; and (3) the lowermost portions of the upper crustal layer, least altered by these processes and thus anticipated to be the purest and least altered anorthosite. 4. Cavity geometry. The impact formed a cavity which underwent proportional growth [Grieve et al., 1981; Spudis and Davis, 1986], whose final diameter approximated that of the outer Rook Mountain ring (about 600 km). 5. Ejecta source. The vast majority of the ejecta was excavated from the 'anorthositic gabbro' upper part of the crust (>90% from depths shallower than 35-40 km) [Spudis et al., 1984; Spudis and Davis, 1986], while the slightly more mafic deepest ejecta (primarily the material of the Montes Rook Formation) may have been derived from the more noritic uppermost part of the lower crust. 6. Paleogeography and ejecta mixing. The distal ejecta impacted onto several pre-orientale mare regions, mixing with the underlying material, and forming locally extensive highalbedo, mafic-rich units overlying the "cryptomaria." 7. Cavity collapse. Collapse of the excavation cavity during and immediately following the cratering event involved the upward movement of the basin floor, the centripetal rotation of the rim into the cavity, formation of a megaterrace by the collapse of the crater rim along a listric fault scarp that is now the Cordillera Mountain ring, and the deformation of the rim deposits into the distinctive texture of the Montes Rook Formation (Figure 17). 8. Basin interior. Cavity collapse was accompanied by inward movement and crenulation of the excavation cavity rim (approximately represented by the outer Rook Mountains), and inward and upward movement of the crater interior along listric faults. The distinctive anorthositic, often highly-shocked inner Rook peak ring [Spudis et al., 1984; Hawke et al., 1991] is interpreted to represent portions of the inner excavation cavity crater wall and floor that have been rotated inward and upward from the deepest, least modified part of the anorthositic upper crust, lying at depths of about 40-50 km (Figure 17). 9. Impact melt generation. Impact melt generated by the event represents a homogeneous mixture of the proximal portions of the target stratigraphy (Figure 18), [Grieve et al., 1991] and was largely derived from upper crustal material (similar in its spectral characteristics to the Hevelius Formation). It lined the inner portion of the cavity and was mobile during cavity modification, ponding in the basin interior. 10. Thermal subsidence. Thermal subsidence followed the short-term modification stage of the basin [Bratt et al., 1985], creating deformation of the melt sheet and accentuating the inner basin depression. 11. Basaltic volcanism. Mare basalt emplacement followed [Greeley et al., this issue] and may have been partly coincident with basin thermal subsidence. 6. ANALYSIS OF PRE-ORIENTALE PALEOGEOGRAPHY (CRYPToMARIA) AND THE EMPLAcEMENT OF ORIENTALEJECTA Background. Prior to the Galileo encounter, there was abundant evidence for pre-orientale mare volcanism in the region of the Schiller and Schickard impact structures southeast of Orientale and south of the Humorum Basin (Plate 1, Figure 3, 19) [Schultz and Spudis, 1979; Hawke and Bell, 1981; Bell and Hawke, 1984] and elsewhere on the Moon [Clark and Hawke, 1991; Robinson et al., 1992]. The Schickard impact structure
17,166 HEAD ET AL.: LUNAR IMPACT BASINS Fig. 18. The relative volumes of melt and clastic debris in impact melt events [after Melosh, 1989; Grieve et al., 1991]. Note that the melt includes the complete stratigraphic column of excavated material and thus its composition is likely to represent an average of the target material. Fig. 19. Airbrush map of a portion of the lunar southwesternear side. The Oftentale basin is in the upper left corner, and Mare Humerum in the upper fight. Locations of profiles in Figure 23 are shown. Arrows indicate positions of Oftentale secondaries that are resolved by SSI data and used in this analysis. contains small patches of relatively fresh mare, indicating that some local post-orientale volcanism has occurred in this area [Schultz and Spudis, 1979; Gaddis and Head, 1981 ]. Analysis of ratio images from the Galileo SSI instrument confirms the strong mafic absorptions of the dark halo craters and the enhanced mafic absorptions in the light plains [Belton et al., 1992b], occurring primarily within the discontinuous deposits of the Hevelius Formation. These data indicate that an extensive region (2-4 x 105 km 2) exhibits a mafic mineral content enhanced over that in typical highlands. It is unlikely that this region represents a zone of enhanced mafic mineral content within the primary ejecta from Orientale, and it is interpreted to indicate mixing of mafic-poor ejecta from the Orientale basin with pre-orientale mare [Bell and Hawke, 1984]. In theory, if the compositional characteristics of ejecta from a basin are sufficiently different from the adjacent substrate on which the ejecta reimpacts, the relationship of the primary ejecta and any excavated secondary ejecta can be studied using mixing models. This is of critical importance in the analysis of models of ejecta emplacement. In practice, however, the ejecta is commonly of sufficiently comparable composition so that this distinction is difficult. However, the presence of cryptomaria (inferred to be mare deposits that predate the emplacement of basin ejecta [Head and Wilson, 1992]) provide a datum of distinctly different composition than typical highland ejecta, and provide a basis for defining the endmembers that are important for mixing models. Mixing between Orientale ejecta, which has a typical mafic-poor highland composition, and pre-orientale mare deposits may occur during the emplacement of Orientale ejecta and secondary impacts by (1) ballistic erosion and sedimentation [Oberbeck, 1975] (Figure 20) or, (2) primarily by postimpact vertical mixing if the ejecta is emplaced by a clustered-impact processes [Schultz and Gault, 1985]. These models make different predictions about the ratio of impactor to excavated mass, and the relative spatial distribution of primary ejecta to excavated material in the ejecta deposits. The ballistic erosion model predicts that the ratio of ejected mass to the mass of the secondary projectile increases logarithmically with radial distance from the primary crater. The ejecta and target material should be well mixed, especially within the continuous ejecta
HEAD ET AL.' LUNAR IMPACT BASINS 17,167 Excavated f ß depos its.. t2 ---.... t3 t4 / -"S L1...' :".::' /._,., r rt.i t. JU u.,,.--,,::;,?.: :.::.":",.. :..:..: ('"' ',:.. : [:.::'"";:?'., / / / -..:!:Y"'"..:½'.':...:.::::::'?'.,;:..½:"'";'"..::?'.: ':,, f f Continuous...'.,:"Seposits ' ½::,...½.":... Ejecta curtain /...".'""'.::'.'.' --- %...:......::.:...::...': :' /..-- --- --.:.' Limit of continuous deposits Lunar surface Vp Vp V x x x x x x x x x xx xx x Xx x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x Primary eiecta Local material Effects of secondary cratering Fig. 20. Ejection process and interaction of ejecta with the underlying lunar surface (target) as a function of radial range [after Oberbeck, 1975]' t, time; Vv, velocity of primary ejecta; Vt +s velocity of mixture of primary and secondary ejecta. During the impact, ejecta leaves 'the cavity along discrete relatively low-angle ballistic trajectories. At any given time these trajectories define an outwardly moving ejecta curtain. Low-velocity, coarse-grained ejecta travel along low-angle trajectories and are concentrated at the base of the curtain, impacting close to the crater and excavating modest amounts of local substrate material (left box). Finer-grained fragments occur higher in the curtain, having been ejected at increasingly higher velocity and steeper angle. As the curtain sweeps radially outward across the substrate, higher velocity ejecta impact the surface, causing increasingly larger proportions of local material to be incorporated in the ejecta deposit. facies. Therefore, in the case of Orientale ejecta interacting with a mare target, the ratio of mare to primary ejecta should increase exponentially with distance from Orientale, and the increase should be smoothly varying. In the clustered-impactor hypothesis, ejecta/target mass ratios are expected to be highly variable depending on impact angle, target strength, cluster size and density, etc. In general, however, the ejecta/target ratios are expected to be a factor of 2-5 less, and for low-impact angles much of the impacting material is expected to ricochet out of the secondary crater. This process is predicted to be most important in isolated impact clusters in the discontinuous ejecta or crater rays. Analysis of Galileo data. The Galileo SSI data provide an ideal opportunity (1) to define the regional extent of pre- Orientale mare deposits in the Schiller-Schickard regions and (2) to test the predictions of models for the emplacement of ejecta in a basin forming event. Spectral mixture analysis (discussed above) was used to deconvolve Galileo SSI data of the region shown in Figure 19 into the relative contributions of the mixed spectral components in the surface material. Careful analysis of inherent spectral variability, and interactive refinement of mixture solutions lead to the selection of three spectrally distinct and geologically meaningful endmembers (Figure 21): mare (Humorurn), highland (Hevelius Formation), and fresh crater (Byrgius). Five-channel Galileo SSI data (relative to the spectral properties of Mare Humorum) for the region shown in Figure 19 were deconvolved into the percent spectral contribution of the three endmembers (Figure 21) using a least squares mixture model. The proportion of each endmemberequired to minimize the error of the fit, subject to the constraint that the sum of the fractional abundances of endmembers equals unity, determines the spectral abundance of that endmember. 'Shown in Figure 22 are the fractional abundance images for the three endmembers and an image of the rms error of the fit. The average fitting error was 1.58 DN, approximately the noise 3OO 250 200 150 100 50 - Galileo Mixing Endmembers ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ' Fresh Crater 0.4 0.5 0.6 0.7 0.8 0.9 1 Wavelength Fig. 21. Galileo five-channel SSI spectra, relative to MH0, of the three spectral endmembers used in the abundance calculations. level of the data. Fitting with additional endmembers causes the least squares solutions to be unstable and result in unrealistic endmember abundances (i.e., >>1.0 or <<0.0). Subtle compositional differences (e.g., TiO 2 abundance) within these major units are not distinguished, but may be inferred from band-residual images [Gillespie et al., 1990] or analysis of ratio images. The error image is generally low across the region with isolated high values associated with fresh craters (which have high spectral variability) and areas where slight misregistration of the spectral bands occurs. These results indicate the three endmembers account for most of the spectral variation in these data. Distribution of cryptomare deposits. Calculated endmember abundances provide a framework for interpreting the relative
17,168 HEAD ET AL.: LUNAR IMPACT BASINS contributions of the surface components represented by the spectral endmembers to the measured multispectral data. The mare fraction image (Figure 22) shows the details of cryptomare distribution; abundance values are density-sliced into discrete levels, with a lower bound of 25% chosen to mark the lower limit of confidence for an unambiguous mare signal (Plate 5). The two zones of high abundance on the floor of the Schickard impact crater correspond to the previously known small patches of post-orientale mare; the rest of the floor shows moderate to low abundance. The east, south, and west walls of Schickard show an absence of mare indicating a surface dominated by highland components. Most of the light plains and discontinuous facies of the Hevelius Formation within and between the Schickard and Schiller craters, and in the northern half of the Schiller-Zucchius basin, also exhibit moderate to low mare abundances. West of the main mare-bearing zone, several small isolated regions of mare are detected which correspond to patches of light plains material within the continuous facies of the Hevelius Formation. The discontinuous patches of "mare" detected south and east of this main zone are artifacts of shading due to the highly variable illumination near the terminator. On the basis of crater counts [Greeley et al., this issue] some of the light plains within Schiller-Zucchius apparently postdate the Hevelius Formation and may be post-orientale maria that have been covered by bright ejecta from nearby craters such as Zucchius. From the mare fraction map in Plate 5, a minimum areal extent of pre-orientale mare is determined to be =3-4 x105 km 2. However, evidence from dark halo craters suggests that additional mare patches are buried by the continuous facies of the Hevelius Formation west of this main area [Schultz and Spudis, 1979; Bell and Hawke, 1984]. These areas are not detected in this analysis (because they have mare abundances less than the 25% cutoff) and may be covered by a greater thickness of ejecta, or be smaller than the spatial resolution of the SSI instrument. The sizes of the cryptomaria outlined here are comparable to several postbasin maria (e.g., Humorurn, Nectaris, Vaporurn, Orientale) and indicate that the geography and geology of the southwest portion of the nearside prior to the Orientale impact was similar to that in the post-orientale period of mare volcanism. Application to basin ejecta emplacement models. Detailed characteristics of mare abundance with radial distance from Orientale provides quantitative information on the mechanism of ejecta emplacement. Nine profiles of mare abundance for the transects indicated in Figure 19 are shown in Figure 23. Profiles A through C indicate the background level of variation for low to zero mare abundance. Profiles D through F cross the southern portion of the cryptomare. The abundance values in these profiles are dominated by typical background levels until 1.5 R where the abundance of mare begins to increase. The mare values rise steadily until they peak, which is then followed by a.'." :... ':'"' 7. :. -. : '":..... 100% 0% i... H F' o F E Fig. 22. Fraction images for the image endmembers m e (top), bighind (upper middle), d fresh crater (lower middle). In the bottom image the s e or of the fit for each pixel analyzed is presented. The average fitting e or was 1.58 DN, approximately the noise level of the data. The concen afions of the t ee surface components e depicted in ese images in shades of gray. High abund ces (100%) e white d low abund ces e gray to black (0%). Since the tot sum of the fraction images must be I %, if ea is high in one endmember (e.g., M e Humorum is high for the m e endmember), then it must be low in the other two. 1F 2F SF F di l Distance From Orient le Fig. 23. Mare abundance as a function of radi distance from the Cordillera Mountains, for the ansects shown in Figure 19.
HEAD ET AL.: LUNAR IMPACT BASINS 17,169 sharp decline to background levels. Profiles F' through H cross the main cryptomare zone. Profile G passes through the post- Orientale mare patches and exhibits greater variability in abundance values. A similar progression to D-F is observed, from low values through steadily increasing values, followed by a steep drop to background levels. All the profiles that cross the cryptomare show a similar trend. Within the cryptomare area, the proportion of mare target material to highland component increases as a function of radial distance from the Orientale basin, and the amount of mare decreases rapidly back to background levels along the distal reaches. This is broadly consistent with the ballistic erosion model for ejecta emplacement. The energy of impact of ejecta and secondary bolides will increase with radial range from the primary impact point, and therefore result in greater excavation of target material. However, most of the cryptomare is within the discontinuous facies of the Hevelius Formation, which becomes thinner and more discontinuous with radial distance thereby decreasing the proportion of highland to mare independently of the mechanism occurring during emplacement of the ejecta. The ballistic erosion model of Oberbeck [1975] and Oberbeck et al. [1975] makes specific predictions for the ratio, g, of mass of ejecta from secondary craters to the mass of the projectile, as a function of range from the primary crater. Pieters et al. [1985] estimated this ratio using telescopic reflectance spectra for secondary craters associated with the crater Copernicus and found the ratios to be consistent with the Oberbeck [1975] model. Several secondary craters from the Orientale event that are resolved by the SSI data are used to provide a similar estimate. The locations of the secondaries are indicated in Figure 19. Radial ranges and secondary crater diameters are determined from Lunar Orbiter photographs. Since it is not possible to determine the precise launch point for the secondary projectiles, three radial ranges are used which bracket all possible launch points. The ranges used are (1) from the center of Orientale, (2) from the Rook mountains, and (3) from the Cordillera mountains. Predicted g factors are calculated from equation (2) of Oberbeck et al. [1975] assuming an impact angle of 75 ø for each of the radial ranges. A factor is determined for each secondary crater from the analysis of SSI data by ratioing the abundance of target (mare) to the abundance of projectile (highland). These values are presented in Table 1. Most of the calculated g values are within a factor of 1.5 of the predicted g value for the three radial ranges, and for some of the secondaries the calculated and predicted values are remarkably close. The calculated g value for secondary A is somewhat lower than the predicted values. Secondary A occurs near the southern wall of Schickard and the low g value may reflect excavation of underlying highland crustal material in addition to the mare surface material during the impact of the secondary projectile. Overall, these results are consistent with the Oberbeck model for ballistic erosion and sedimentation during emplacement of ejecta in the distal regions beyond the continuous ejecta. It is not possible to compare the predictions of the Oberbeck model with those of Schultz and Gault [1985] since they make similar predictions for isolated secondaries such as those investigated in this analysis. The spatial and spectral resolution of the Galileo SSI data is not sufficient to separate the effects of the emplacement of the Orientale ejecta from post-orientale modifications of the surface for secondary crater chains etc. Detailed comparison of these models must await the acquisition of high spatial resolution multispectral or Secondary Crater A TABLE 1. Orientale Ejecta gt Factors Radial Crater Ran e*, Diameter Predicted Calculated 1219 20 1.81 1.09 950 20 1.61 1.09 760 20 1.43 1.09 1406 18 2.01 1.68 1137 18 1.83 1.68 947 18 1.68 1.68 1486 8 2.84 2.83 1217 8 2.61 2.83 1027 8 2.41 2.83 1513 15 2.23 2.54 1244 15 2.05 2.54 1054 15 1.91 2.54 E 1674 8 2.98 3.27 1405 8 2.78 3.27 1215 8 2.61 3.27 All gt factors calculated assuming an impact angle of 75 ø. *The three radial ranges correspond to the distance from the center of Orientale, the radius of the Rook Mountains, and the Montes Cordillera. hyperspectral data over a target area like the Schiller-Schickard cryptomare. 7. CHARACTERISTICS AND SPECTRAL PROPERTIES OF SOUTH POLE- AITKEN AND OTHER SOUTHEASTERN FAR SIDE BASINS Background. The Galileo data revealed a major anomaly in the color-ratio composite images on the southern far side, associated with the South Pole-Aitken basin region [Belton et al., 1992b] (Figure 3 and Plate 1). This pre-nectarian (>3.92 Ga ago) impact basin is extremely large (>2000 km; Figure 24) and is so degraded that it went undetected for many years. Only the proposed near-side Procellarum basin is larger (-3200 km) and older [Wilhelms et al., 1979]. The presence of a large farside basin was originally predicted on the basis of the presence of mountains near the near-side south pole [Hartmann and Kuiper, 1962], and additional evidence for far-side mountain ranges was revealed by Apollo [Wilhelms et al., 1969]. However, it was not until Zond images [Rodionov et al., 1971; Kislyuk, 1975; Rodionov et al., 1976; Shevchenko, 1980; Shpekin, 1983] and Apollo laser altimetry [Wollenhaupt and Sjogren, 1972a,b; Kaula et al., 1973] demonstrated the presence of a huge depression (about 5-7 km deep; Figure 25) that the previous observations were integrated into the concept of a major impact basin [Howard et al., 1974; Stuart-Alexander, 1978; Wilhelms et al., 1979; Leikin and Sanovich, 1985]. Due to the extremely large size and degraded nature of the basin, the exact diameter of the structure and the location of its rings are a matter of debate. Stuart-Alexander [1978] proposed that the basin is 2000 km in diameter and centered at -50 ø, 180 ø, while Wilhelms et al. [1979] map a basin 2500 km in diameter and centered at -56 ø, 180 ø, and a possible inner ring 1800-2000 km in diameter is proposed by Wilhelms et al. [1979] (Figure 24). Wood and Gifford [1980a] mapped 28 possible rim segments and estimated a diameter of about 2600 km with a center at-60 ø, 180 ø. They point out, however, that the rim suggested by the altimetry [Wollenhaupt et al., 1973] is 300-600 km larger than the rim they define morphologically. Finally, Leikin and Sanovich [1985] used Zond 8 data to yield a N-S profile at about -150 ø, and combined
17,170 HEAD ET AL.: LUNAR IMPACT BASINS 90 ø 120 ø 150 ø 180 ø 210 ø 240 ø 270 ø 0ø I...,,,... t t0ø TsioIko sky. _, o? o. :.,.; ;. / / A k n Basin,- / ', 30 o 30 ø -.,. h edinger ß o 60 ø. 60 South 90 B Pole A [4 South Pole-Aitken Basin l A' 140 160 180 200 220 6 B South Pole-Aitken Basin B' Fig. 24. Distribution of rings and basin centers (numbers) defining the South Pole-Aitken basin according to several hypotheses. 1, (-60 ø, 180 ø) [Wood and Gifford, 1980a]; 2, (-56 ø, 180 ø) [Wilhelrns et al., 1979]' 3, (-50 ø, 180 ø ) [Stuart-Alexander, 1978]; 4, (-41.5 ø, 176.5 ø ) [Leikin and $anovich, 1985]. Outer circle is edge of Moon. Outer ring is from Wilhelrns et al. [1979] and the inner ring is from Stuart- Alexander [1978]. Basins are as follows: (AU, Australe; AP, Apollo; AG, Amundsen-Ganswindt; AN, Antoniadi; I, Ingenii; K, Keeler- Heaviside; PO, Poincar6; PL, Planck; SC, Schr/3dinger; SR, Sikorsky- Rittenhouse. Rings are circles around centers; see references for detailed ring locations. 4 2 km 0-2 -4-6 '? ' ' ' '.;o Fig. 25. Topographic profiles for the rim and interior of the South Pole-Aitken basin (data from Kaula et al. [1973], E-W profile; Leikin and $anovich [1985], N-S profile). viscosities during the period (pre-nectarian) when viscous processes were important; a difference of 50-100øC in the upper crust would be enough to explain difference in effective viscosities. The South Pole-Aitken basin is also a major negative free air gravity anomaly and, Apollo, lying within the basin, is the largest negative anomaly on Moon [Ferrari, 1977]. A number of large craters and basins are superposed within with Apollo 15 and 17 altimetry, determined a basin diameter of 2200 km centered on -41.5 ø, 183.5 ø. As a working model, we adopt here the two diameter estimates and center positions proposed by Stuart-Alexander [1978] and Wilhelms et al. [1979]. The location of these rims is shown in Figure 25. Further evidence for the presence of the basin [Wood and the basin (e.g., Apollo, Ingenii, Schr/3dinger, Planck) or on the rim (e.g., Keeler-Heaviside, Australe), commonly obscuring the continuity of the rim structure (Figures 24 and Gifford, 1980a] is found in the paucity of craters >25 km in 26). These younger basins have clearly deposited ejecta within diameter in the interior (18/380,000 km 2) relative to the South Pole-Aitken, or excavated its floor (Figure 26). For exterior (25/380,000 km 2) [Wood and Gifford, 1980b], and in example, the Nectarian-aged Ingenii (325-km diameter) and the the concentration of mare deposits (Figure 26) and associated pre-nectarian aged Apollo (503-km diameter) basins excavated volcanic features [Scott et al., 1977], comparable to other farside regions. Virtually all mare deposits occur within post- South Pole-Aitken craters and basins in the northern half of the basin (Figure 26), and are mapped as Imbrian in age [Stuartmaterial from the basin floor and deposited in the interior and on the rim. In the southwestern part of the basin, the 312-kmdiameter Imbrian-aged peak-ring basin Schr/3dinger has deposited ejecta over most of this quadrant, and has obscured Alexander, 1978; Wilhelms et al., 1979]. Although no two previously formed basins of comparable size (Figure 26). cryptomaria have been previously mapped in the South Pole- Aitken basin, the presence of widespread patches of later mare, and the great depth of the basin, strongly suggesthat they may be present [Head and Wilson, 1992]. The South Pole-Aitken basin appears to be anomalously deep (up to 7-8 km below the adjacent highlands; Figure 25 for a basin of its size and age. Viscous relaxation of basin topography, particularly in the early history of the Moon when the thermal gradient was steeper, together with mare infilling serve to reduce the initial topography of impact basins [Solomon et al., 1982]. Many older near-side basins (e.g., In summary, the stratigraphic sequence for the region consists of pre-south Pole-Aitken crust (the target material), the deposits of the basin itself (see Figure 27 for an estimation of the distribution of these if scaled from Orientale), subsequent craters and basins and their internal and external deposits of pre-nectarian through Imbrian age, possible cryptomaria deposits, patchy Imbrian-aged mare plains, and later Eratosthenian and Copernican craters. Orbital gamma ray spectrometry data are available for the northern part of the basin. Total lunar radioactivity on the central far side is generally low except for an anomaly Tranquillitatis, Fecunditatis) have lost their long-wavelength extending from 160øE to 166øW, the northern part of the topographic expression. Solomon et al. [1982] attribute the South Pole-Aitken basin [Metzger et al., 1973a,b; Trornbka et depth of South Pole-Aitken and its difference from the near-side al., 1973]. Although mare patches exist under the ground track, basins to a factor of 10 difference in near side-far side crustal they do not appear extensive enough to account for the
HEAD ET AL.' LUNAR IMPACT BASINS 17,171 Fig. 26a. Figures 26a-26c show mare infilling of far side and limb basins. Distribution of maria within the South Pole-Aitken basin [from Stuart-Alexander, 1978; Wilhelms et al., 1979; Wilhelms, 1987; Wilhelms and El Baz, 1977]. Also shown is the location of post-south Pole-Aitken basins (see Figure 24 for details of basins). anomaly. In addition, the highest concentrations of Fe and the lowest concentration of Ti (too low for typical maria) occur within the basin, in the vicinity of the crater Van de Graaff. The concentration of Mg is about average within the basin. The Van de Graaff region within the basin is unique in the far-side orbital geochemical data: values for Fe, Th + U, and K are higher than typical terra and lower than most maria; Mg and Ti are lower than typical maria. Thus, despite the differences in the interpreted location of the basin ring, the topography and other characteristics of the southeastern far side clearly indicate the presence of a basin of the order of 2000-2500 km diameter, over twice the size of the Orientale basin. The distribution of the massifs, proposed rings, maria, and topography (Figures 24-27) permit several plausible models for the ring and interior structure. While recognizing that further analysis is required, we tentatively adopt the position of Solomon et al. [1982] that the average 2000 or 2500 km diameter value is equivalent to the Orientale Cordillera ring, and we map out the equivalent deposits in the interior of the basin on the assumption that the processe scale to this dimension. We use this, and the maps of subsequent events (Figures 27, 25, 26) to aid in the interpretation of Galileo data. In particular, styles of mare fill of very degraded basins such as Smythii [Wilhelms and EI-Baz, 1977] and Australe [Whitford-Stark, 1979] may be helpful in the interpretation of the South Pole-Aitken deposits (Figure 26). Initial analysis of Galileo data showed that the interior of the basin is characterized by low albedo and distinctive color properties relative to the surrounding highlands (low 0.41/0.76 lam; enhanced mafic absorption relative to typical highlands). Areas in and around the basin require special consideration due to observation geometry; error analyses are discussed more thoroughly in McEwen et al. [this issue] and Pieters et al. [this issue]. The unusual properties of the basin were interpreted by Belton et al. [1992b] to indicate that mafic components are enhanced in the generally feldspathic highland materials of the basin. Thus, the South Pole-Aitken basin interior is not typical feldspathic highland crust. The lowalbedo feature seen by Galileo within the basin has a diameter of about 1400 km and is centered at about -47 ø, 170 ø (Figures 5 and 28 and Plate 6). We now examine the Galileo spectral parameter maps and spectral unit classification maps of the South Pole-Aitken basin and the surrounding region to assess further these characteristics. Results from spectral parameterization. Spectral parameter maps of the southeastern far side are shown in Figure 13. Both visible spectral slope and estimated depth of the mafic absorption band exhibit heterogeneities correlated with some, but not all, basin materials of comparable albedo. The interior of Mendel-Rydberg exhibits both a less red visible spectral slope and a stronger mafic band than the surrounding Hevelius. As a whole, the interior of Hertzsprung is less mafic than its exterior. Hausen and its ejecta exhibit a redder visible spectral slope and a slightly stronger mafic band than the distal Hevelius Formation to the northeast. On a larger scale, there is a perceptible variation in mafic absorption band depth in the
17,172 HEAD ET AL.: LUNAR IMPACT BASINS Fig. 26b. Mare Smythii (Apollo 15 photograph AS15-12992). This view shows the 360-km-diameter interior ring which is partially filled with maria, Imbrium/Nectarian furrowed and pitted basin floor material, and crater ejecta deposits [Wilhelrns and El-Baz, 1977]. Evidence also exists for cryptomaria in the vicinity of Mare Smythii [Clark and Hawke, 1991]. The outer basin ring, 840 km in diameter, lies outside this image. 8'0 9'0 10"0 11'0 ' 60-70 8 I1 100 110 I i i I Fig. 26c. Sketch map of the distribution of mare deposits in Mare Australe; the dashed line represents the basin rim, about 880 km in diameter [modified from Whitford-Stark, 1979]. Mare Australe and Smythii may be similar in appearance to the interior of the South Pole- Aitken basin prior to the formation of Orientale and other large nearby basins. 30- interbasin terra. Band depth increases from northeastern portions of the region to the south and southwest, up to the rim of South Pole-Aitken, where its depth increases abruptly in the basin interior. Both the interior of South Pole-Aitken basin and the superposed Apollo basin exhibit spectral properties distinct from surrounding highland regions [Belton et al., 1992b]. Most obvious is the overall lower albedo (Figures 5 and 28 and Plate 6). Lowest albedos are found within Apollo and the northcentral part of South Pole-Aitken; southeastern South Pole- Aitken is higher in albedo, though not as bright as the surrounding highlands. Less obvious is that bright fresh craters, which are prevalent in surrounding highlands observed at comparable illumination geometry, are almost entirely absent from the interiors of these two basins. The spectral parameter images in Figure 13 show that the basin interiors also have a redder visible spectral slope than any adjacent highlands, including the proximal Hevelius Formation, and a stronger mafic absorption than the cryptomaria, comparable to that in the mare. Unlike the cryptomaria, however, the relationship of albedo and mafic band depth is different. Within South Pole-Aitken, it is the brighter southeastern portion which has the strongest mafic absorption and the reddest visible spectral slope. In contrast, in the cryptomaria where mare are believed to be mixed with mafic-poor higher-albedo material, lower-albedo regions with the largest fraction of intermixed mare have the strongest mafic absorption.
HEAD ET AL.: LUNAR IMPACT BASINS 17,173 Fig. 27. Inferred distribution of basin deposits within the South Pole-Aitken basin on the basis of scaling from Orientale and interpretation of the outer 2500-km-diametering as equivalen to the Montes Cordillera. This initial analysis suggests the presence of at least two types of materials in Apollo and South Pole-Aitken: (1) lowalbedo mafic material in Apollo and north-central South Pole- Aitken and (2) higher-albedo, more mafic material in southeastern South Pole-Aitken (Figure 29). The fundamental differences in their spectral properties from those of other areas suggesthe presence of different materials than in other parts of the southeastern far side and the western limb. appears comparable to that in Schickard crater (-200 km in diameter). The southeastern part of Hertzsprung has been buried by the distal Hevelius Formation, but the exposed northwestern part of the basin exhibits a remarkably similar pattern of spectral units to that in Orientale. The largest-scale heterogeneity is that the basin interior as a whole is less mafic than the exterior ejecta deposits. In detail, a central smooth patch located Compositional interpretation of these spectral properties is addressed in detail by Pieters et al. [this issue]. Superficially, however, the low albedo, red spectral slope, and strong mafic absorption in the low-albedo unit resemble the properties of some mare, and the higher albedo and strong mafic absorption of the southeastern deposits is reminiscent of the Schickard cryptomare. For these reasons, spectral classification models may inaccurately classify the basin materials with mare or similarly to the Maunder Formation and the inner basin ring classify spectrally with the mafic-poor interior deposits and proximal ejecta of Orientale. Spectra of these units are shown in Figure 14, compared with the corresponding Orientale deposits; the major differences are in continuum slope. On the mean-standard deviation plot, these Hertzsprung units are separated from each other and from the corresponding Oftentale deposits in a direction consistent with differences in regolith cryptomare, even though they are clearly different. maturation rather than lithology. However, most of the basin Nevertheless, because of their distinctiveness, mixing of the South Pole-Aitken and Apollo materials with other highlands may still be discerned by these models. Results from spectral unit classification. In areas outside South Pole-Aitken and Apollo, the spatial distribution of spectral heterogeneitie shown in the unit map in Plate 4 and interior classifies with with the Montes Rook Formation, and it also exhibits a similar spectrum. Unfortunately, the central smooth patch and inner ring are too small and insufficiently sampled for a meaningful comparison of the mafic band depths of basin interior materials. The surrounding exterior deposits classify with the distal Hevelius Formation, and are similarly their spectral character shown in Figure 14 are consistent with more mafic than the basin interior. Their less-red continuum heterogeneities evident in the spectral parameter maps. For example, the inner ring of Mendel-Rydberg is largely filled with "higher-albedo mafic material" and has a local concentration of "fresh mafic craters" resembling those in the maria. The spatial extent of the possible cryptomare deposit slope is attributable to young craters, including Vavilov whose bright ejecta deposits cover much of the region. Hausen is distinguished by a patch of material occupying parts of the interior and an annulus up to about one crater diameter in width, classifying distinctly from the surrounding
Plate 5. Mare abundance determined from spectral mixture deconvolution of Galileo SSI data. Gray and white represent <25% mare, and purple represents 100% mare. The intermediate level abundance intervals between these extremes are shown in the color bar. Plate 6. Albedo map of South Pole-Aitken basin interior. Unit brightnesses are defined relative to MH0 (Mare Humorurn 0) [see Pieters et al., this issue]. White (<3.10 MH0); light pink (2.45-3.10 MH0); medium pink (1.85-2.45 MH0); dark purple (1.55-1.85 MH0); black (<1.55 MH0). Note that the low albedo mafic deposits are concentrated in areas where mare patches are abundant and where extensive mare deposits are seen in Orientale (compare to Figure 27), Smythii and Australe (compare to Figure 26). Note also that the higher-albedo mafic deposits are concentrated in the southeastern part of the interior, equivalent to the position of the Montes Rook Formation in Orientale (compare to Figure 27), and to a position within South Pole-Aitken where few younger basins are present (compare to Figure 24).
HEAD ET AL.: LUNAR IMPACT BASINS 17,175 Fig. 28. Shaded relief map of South Pole-Aitken basin interior. distal Hevelius Formation. The spectrum in Figure 14 shows three major distinctions from the surrounding distal Hevelius Formation: a redder visible spectral slope, a stronger mafic absorption, and a shorter-wavelength shoulder of the mafic absorption. As discussed by Pieters et al. [this issue], all three of these attributes also distinguish South Pole-Aitken from the surrounding highlands. This evidence suggests an affinity of the material composing Hausen with the nearby interior of South Pole-Aitken. In contrast to Hausen, nearby Bailey is undistinguished from its surroundings. The low-albedo and high-albedo units within South Pole- Aitken and Apollo classify as low- and higher-albedo mafic materials and fresh mafic crater deposits. As explained earlier, this grouping can be attributed to the superficial spectral similarity rather than to a genetic similarity. However, the classification of the Nectarian terra between basin ejecta deposits does show a clear relationship to distance from South Pole-Aitken and Apollo. In the northeast portion of Plate 4, the Nectarian terra classify as distal Hevelius Formation plus a component of "high-albedo mafic material." Moving southwest, the component of "high-albedo mafic material" increases until it dominates an annulus outside the rim of South Pole-Aitken. The annulus is widest and best developed adjacent to Apollo. This trend is identical to the southwestward increase in mafic band strength in the Nectarian terra, noted above in discussion of the spectral parameter images. CENTER South Pole-Aitken [ Basin Rim.. ' Apollo Basin Cryptomare,,=.,,,= =,.,=' Megaregolith '., Maria _ Crater EJecta...,'... SE Fig. 29a. Schemati cross section showing the interpreted three-dimensional structure and stratigraphy of the South Pole- Aitken basin. Portions of the basin interior are severely modified by large craters and basins, such as Apollo. Mare flooding may have begun in pre-imbrium times, with early deposits covered and modified by subsequent events. The southeast portion of the basin is least modified by subsequent impact basins and may contain the deepest and least modified basin ejecta (equivalento the Montes Rook Formation in Orientale).
17,176 HEAD ET AL.' LUNAR IMPACT BASINS \\ i I Fig. 29b. Paleogeographic maps illustrating the interpreted sequence of events (2500-km-diameter outer ring is used). Upper left is immediately following the basin event. Upper fight is events occurring dufing the later pre-nectarian pefiod showing the major basins (ring is topographic rim except in the case of Apollo at the upper right where inferred ejecta deposits are shown); initial mare deposition to produce cryptomaria may have begun as early as this time. Lower left is major events occurring in the Early Imbrian (Lower Imbrium Series), showing Schr/Jdinger and Orientale and their ejecta. Lower fight is airbrush map of present configuration. 8. INTERPRETATION OF THE SOUTH POLE-AITKEN BASIN On the basis of these analyses and those of Pieters et al. [this issue], there appear to be two types of material associated with the South Pole-Aitken Basin interior and predating the mapped mare basalt patches shown in Figure 26. Together, these materials have characteristics that are in contrast to much of the highland material outside the basin, which typically is higher albedo, less red, less mafic, and has a shoulder to the l- gm mafic mineral absorption located at 0.76 gm. One of the types of material is located in the northern and central part of the basin and is characterized by the lowest albedo and a mafic absorption band as strong as typical 1ow-Ti mare soils. The second type of material is concentrated in the southern part of the basin and is characterized by a higher albedo than the material to the north (although still lower than the typical surrounding highlands), a stronger 1-gm absorption than the northern unit, and a shorter wavelength shoulder to the 1 gm absorption (near 0.56 tm) (Figure 28). The fundamental
HEAD ET AL.: LUNAR IMPACT BASINS 17,177 differences in these spectral properties from those of other areas suggest the presence of different materials than in other parts of the southeastern far side and the western limb [Pieters et al., 1992]. Compositional interpretation of these spectral properties is addressed in detail by Pieters et al. [this issue], who propose several possible hypotheses that are consistent with the unique spectral reflectance properties of the interior unitse One is the southern interior region. The spectral properties of this region could be explained by the presence of (1) ferric pyroxene, (2) high-iron, low-titanium homogeneous glass, (3) olivine [Pieters, 1982], and (4) unknown lithologies. They conclude that options 2 and 3 are most likely. Another hypothesis basin, the region representing the melt sheet should be more mafic than the average upper crustal mature highlands (having sampled and averaged both upper, and lower crustal, more mafic, material), but perhaps less mafic than the immediately adjacent ejecta deposits representing the most deeply excavated material (Figure 29). However, the coherency of such a relatively thin impact melt unit (estimated to be about 1 km in Orientale [Head, 1974a]), would not seem capable of surviving the subsequent formation of hundreds of craters and basins in the basin interior, and the emplacement of maria (Figures 26 and 29). These subsequent impacts would have excavated primarily deeper crustal material or perhaps upper mantle material depending on the size of the impact and the thickness concerns the north central interior region. The albedo and 1- tm of the post-south Pole-Aitken crust. In contrast, the band character of this region could simply be explained by the presence of low-ti mare basalt soils, without recourse to olivine or glass contributions. However, the low spectral and spatial resolution SSI data mean that marelike characteristics are not uniquely determined and this leaves open the possibility of (1) small mare patches, (2) cryptomare, (3) soil formed on mafic lower crust excavated by the basin-forming event, (4) soil developed on Fe-rich impact melt mixed with compositionally anomalous southern interior region of the basin lies in the area that is a candidate for the deposits of the collapsed basin rim (the equivalent of the Montes Rook Formation in Orientale; Figures 27 and 29). Here, there is a paucity of post-south Pole-Aitken basins (Figures 24 and 29). In this case, the material might be most plausibly interpreted as containing olivine excavated from the lowermost crust or uppermost mantle. feldspathic lithologies, and (5) some combination of the How then to account for the relatively homogeneous lowabove. Pieters et al. [this issue] conclude that soils developed albedo northern interior deposits? The most plausible on the central portion cannot be distinguished from low-ti explanation appears to be a combination of (1) deep crustal mare basalts and that the southern part of the basin contains material exposed in basin interior massifs and in the rims and soils that appear to be enriched in olivine or Fe-rich, Ti-poor deposits of subsequent craters, and (2) admixed basaltic homogeneous glass. material from early lava filling, partly buried by superposed The nature and distribution of structures, morphological crater and basin (e.g., Orientale) ejecta (cryptomaria). The deposits, and spectral units in less complex regions may help presence of cryptomaria is logically to be inferred from (1) the in the interpretation of these data. The reconstruction of the concentration of mare patches in the same region as the low- South Pole-Aitken basin interior using the Orientale analogy albedo anomaly (Figures 26 and 28 and Plate 6), (2) the spectral (Figure 27) and the distribution of maria in degraded basins reflectance characteristics of this region (most simply such as Australe and Smythii (Figure 26) provide a framework interpreted as low-ti basalt soils), (3) the similarity of the for this discussion. Two classes of events, formational and position of the low-albedo region with mare fill in Orientale modificational, have been important in the history of the basin (compare Figures 27 and 28 and Plate 6), and (4) the similarity (Figure 29). The impact event itself was accompanied by of styles of mare fill in large degraded basins such as Smythii excavation deep into the crust and perhaps upper mantle, and Australe (Figure 26). On the basis of analyses of mare fill collapse of the cavity to form multiple rings, and emplacement in Australe [Whitford-Stark, 1979], the average thickness could of a melt sheet. Modification of the basin included the be of the order of 750 m, not a significant decrease to the basin interior topography. In this case, the many small mare patches observed today are simply the last deposits of what was originally a much more extensive mare filling (such as that seen in Mare Australe and Smythii; Figure 26) which is now formation of subsequent craters and basins (Figures 26 and 29) and the emplacement of volcanic deposits, events which are likely to have occurred simultaneously. Distribution of mafic deposits on the northeast rim of the basin (Figure 13 and Plate 4), as well as in its interior, strongly suggests that at least some component of the enhanced mafic mineral content is due to the excavation of material from depth, either by the basinforming event itself and/or subsequent events such as the formation of the Apollo basin (Figure 29). In either case, the widespread continuous nature of the enhanced mafic content seems unlikely to be due to the presence of either primary surface basalt units or ejecta derived from subsequent impacts into a thin basalt cover in the basin interior. The orbital geochemical data for iron and radioactive elements [Metzger, 1974] for the northern rim and interior of the basin, although anomalous for the far side in general, are not in the range anticipated for the lunar mantle and are most consistent with crustal materials. Impact melt deposits were certainly a major factor in the basin interior (Figures 27 and 29) in its early history, and compositionally they should be representative of the average composition of the target material, not the deepest material [Grieve, 1991; Grieve et al., 1991]. Therefore, because of the extremely large size of the largely buried by ejecta and revealed only by its anomalous albedo and spectral characteristics (Figure 29). The anomalies observed in the Galileo data could thus be explained by a combination of these factors. The concentration of low albedo, relatively mafic materials within the innermost part of the basin, together with the abundant small patches of mare material in the same area, would seem to favor the cryptornate hypothesis. However, the spectral properties of the region also appear locally to extend beyond the basin rim (in the east), and this observation suggests that mafic material might have been ejected from the basin during its formation, or from subsequent basins such as Apollo. The southern basin anomaly may be related to deep South Pole-Aitken ejecta largely unmodified by subsequent major events (Figure 29). Higher spatial and spectral resolution spectrometer data, as well as complete geochemical coverage of the basin is necessary to confidently distinguish further between these hypotheses.
17,178 HEAD ET AL.: LUNAR IMPACr BASINS 9. SEQUENCE OF EMPLACEMENT OF BASIN MATERIALS ON THE $OLrrHEASTERN FAR SIDE The relationships of the spectral properties of southeastern far-side basins with surficial geology suggests that specific basin deposits can be isolated and their spectral properties analyzed using Galileo data. Consequently, aspects of the geologic history of these basins can be inferred by relating spectral properties to physical geology revealed by higherresolution Lunar Orbiter, Zond, and Apollo images. The physical geology of the region has been discussed by Scott et al. [1977], Stuart-Alexander [1978], Wilhelrns et al. [1979], and Wilhelrns [1987], and their observations and interpretations are incorporated into the discussion below. The earliest major basin to form in this region was South- Pole Aitken, which exposed mafic-rich materials distinct from the relatively mafic-poor surrounding highlands. As evidenced by the mafic-rich annulus of the basin, some of its interior material escaped the basin as ejecta either at the time of formation of South-Pole Aitken or due to subsequent large impacts. The exterior South Pole-Aitken ejecta were mixed with the underlying crust by prolonged cratering and superposition of overlapping ejecta blankets of smaller basins, to the extent that the South Pole-Aitken ejecta is now recognized as an increase in mafic band depth approaching the basin rim. Later in pre-nectarian time Apollo was superposed on the northeast part of South Pole-Aitken's interior. Two important observations about the spectral properties of the interior of Apollo and the surrounding regions allow inferences to be made about the composition of the crust here. First, Apollo does not expose a material that is very distinct from that in the rest of north-central South Pole-Aitken. Second, the relatively mafic annulus around South Pole-Aitken is widest and best developed adjacent to Apollo. The similarity of Apollo to the rest of north-central South Pole-Aitken suggests that the low-albedo material either continues to depth, so that Apollo excavated more of the same, or was superposed on the interiors of both basins. The wider higher-albedo mafic annulus adjacent to Apollo is consistent with emplacement of Apollo ejecta outside the rim of South Pole-Aitken. If this interpretation is correct, it implies that the low-albedo material was excavated by Apollo and continues to depth. In this case it may represent lower crust material, or a derivative thereof. In later pre-nectarian and Nectarian time, Hertzsprung, Mendel-Rydberg, and Bailly were formed in distal parts of the South Pole-Aitken ejecta. Hertzsprung's interior is less mafic than that of South Pole-Aitken, having been derived from shallower, less mafic crustal layers. The interior forms a "window" into the less mafic crust underlying ejecta of Apollo and South Pole-Aitken, but the exterior ejecta are slightly more mafic than the basin interior due to intermixing with the preexisting ejecta during either emplacement or subsequent cratering. After these basins formed, some mare-type volcanism is inferred to have occurred. Mare deposits partially filled Mendel-Rydberg as well as Schickard and Schiller- Zucchius, and probably the interior of South Pole-Aitken. The youngest basin to form in this region, Orientale, is so large that it dominates surficial geology of the southeastern farside and western limb (Figure 2). Shallowly derived, maficpoor material comprising the Hevelius Formation formed the exterior ejecta deposits. These covered and intermixed with preexisting materials, whose spectral properties somewhat modified the Hevelius. The marelike deposits in Mendel- Rydberg, Schickard, and Schiller-Zucchius imparted a significant mafic component to the overlying Hevelius, especially where subsequent small craters excavated local concentrations of the mare component. Hevelius superposed on the surfaces east of Orientale, proximal to the basin, retained a very mafic-poor lithology. Hevelius superposed on surfaces to the west incorporated a larger mafic fraction from the preexisting surface including primarily South Pole-Aitken ejecta, as well as impact melt sheets and possibly unrecognized cryptomaria, during either emplacement or subsequent cratering. Orientale ejecta also buried Bailly and the southeastern half of the interior and ejecta of Hertzsprung, but because of their subtle spectral differences they did not modify the Hevelius enough to be detected by Galileo images. The interior of Orientale exposes material with spectral properties similar to that exposed in the interior of Hertzsprung, mostly less mafic than the surrounding terra. However, a partial annulus between the Cordillera and Inner Rook Mountains is comparably mafic to the exterior deposits and probably represents the most deeply derived ejecta. The incomplete correlation of the annulus with the Montes Rook Formation implies that the enhanced mafic content and the distinctive texture may have different origins. After Orientale formed, minor mare volcanism continued and Mare Orientale was emplaced, as well as small patches superposed on the cryptomaria in Mendel-Rydberg and Schickard [Greeley et al., this issue]. In Eratosthenian time the crater Hausen formed, and penetrated the distal Hevelius Formation to expose a redder and more mafic material, perhaps a thick accumulation of ejecta from South Pole-Aitken just outside that basin rim. 10. CONCLUSIONS The multispectral image data acquired by the Galileo SSI experiment during the December 1990 encounter provide important information on the composition of the western lunar limb and parts of the far side and the nature of impact basins, their deposits, and stratigraphy. Spectral and statistical parameterization, mixing model analyses, and spectra derived from the multispectral image data were compared to photogeologic analyses of higher resolution images. The ejecta deposit of the Orientale basin, the Hevelius Formation, appears relatively homogeneous and similar in composition to mature Apollo 16 soils, and is interpreted to be excavated from the upper crust. The centrally located Maunder Formation, although giving a crater age slightly younger than the Hevelius, is distinct from the younger mare basalts and does not show a pronounced mafic absorption band, being comparable to the Hevelius Formation in its spectral reflectance properties. These characteristics support the interpretation of the Maunder Formation as impact melt, representing the average composition of highland upper crustal target material. The Montes Rook Formation, located in an annulus between the Maunder and the Hevelius shows a slightly enhanced mafic absorption and may incorporate deeper crustal material. The distal Hevelius Formation and associate deposits show local mafic enhancements, the largest of which are interpreted to represent the presence of pre-orientale mare deposits, or cryptomaria. The characteristics and sizes of two of these, Schiller-Schickard and Mendel-Rydberg, show that the emplacement of basalts in ancient impact basins to produce maria of sizes comparable to those presently observed, was
HEAD ET AL.: LUNAR IMPACT BASINS 17,179 widespread in this region. Using the Schiller-Schickard cryptomare deposits as a datum, mixing-model analyses were employed to estimate the proportions of Orientale ejecta that were mixed with the underlying mare deposits as a function of range from the basin center. The results are consistent with the ballistic erosion and sedimentation model for ejecta emplacement in the distal regions beyond the continuous ejecta deposit. One of the most fundamental discoveries from these data is the presence of a huge compositional anomaly on the southern lunar far side. This low-albedo region is much more mafic than the surrounding highlands and corresponds to the interior of the pre-nectarian South Pole-Aitken impact basin, about 2000-2500 km in diameter. In addition, the proportion of mafic material in the highland soils increases as the basin is approached. Within the basin, distinct variations in the albedo and spectral reflectance properties are also observed. The South Pole-Aitken basin compositional anomaly is interpreted to be due to several factors. The large size of the basin (over twice the diameter of Orientale) and its spectral reflectance characteristic suggest that it excavated to deeper levels than Orientale, into the more mafic lower crust. There is no positive evidence that the deposits represent material excavated from the upper mantle, although some contribution from the mantle cannot be ruled out. The nature and distribution of the deposits suggesthat some components of the mafic anomaly represent ejecta: (1) deep ejecta from the South Pole-Aitken basin (e.g., the possible presence of olivine in the southeastern interior), and (2) ejecta on the basin rim from South Pole-Aitken or subsequent basins such as Apollo. The spectral reflectance properties of much of the interior are consistent with and most plausibly interpreted as early volcanic fill (cryptomare), similar to that seen in ancient basins such as Smythii and Australe. Subsequent impact events both inside and outside the basin (such as Orientale) mixed nonmare material with these deposits. In Imbrian times, emplacement of mare deposits continued in the form of emplacement of numerous small patches in the basin interior. Examination of the deposits of these impact basins on the western near-side limb and parts of the far side show that although basin-forming events are an important factor in producing lateral heterogeneities in crustal composition, and in modifying preexisting deposits (such as cryptomaria), the majority of material excavated is confined to crustal levels. The results of this study are consistent with a gradational vertical crustal stratigraphy consisting of an uppermost crustal layer of anorthosite, mixed basin ejecta, and cryptomaria deposits (generally corresponding to the megaregolith), an upper crustal layer of anorthosite, and a lower more mafic noritic layer. Examination of the deposits of the proposed Procellarum basin (about 3200-km diameter [Whitaker, 1981])on the second Galileo lunar encounter (December 1992) may provide more information on the deep excavation of material by impact basins. At the present time, continuing incomplete knowledge of the details of crustal thicknesses and stratigraphy, the mode of impact basin cavity growth and source of ejecta as a function of depth, and the comprehensive compositional characteristics of the lunar surface preclude more refined treatment of these problems. These data, in spite of their limited spectral and spatial resolution, do however show how important compositional data are for the understanding of basin formation processes and crustal stratigraphy. Many of the basic questions remaining from this study could be addressed by global high-resolution geochemical and mineralogical data, such as could be obtained by polar orbiting spacecraft. Acknowledgements. We wish to gratefully acknowledge the members of the Galileo Team at the Jet Propulsion Laboratory for their performance in mission planning and data acquisition during this encounter. This work was supported by NASA contract 958512 from JPL to J.W.H. and by NASA grant NAGW-713 to J.W.H. from the Planetary Geology and Geophysics Program of the Solar System Exploration Division of NASA. The help of Erich Fischer, Joel Plutchak, Jessica Sunshine, and Karen Plouff is gratefully acknowledged. Richard Grieve and Don Wilhelms provided detailed and helpful reviews. REFERENCES Adams, J. B., M.O. Smith, and P.E. Johnson, Spectral mixture modeling: A new analysis of rock and soil types at Viking Lander 1, J. Geophys. Res., 91, 8113-8125, 1986. Bell, J., and B. Hawke, Lunar dark-haloed impact craters: Origin and implications for early mare volcanism, J. Geophys. Res., 89, 6899-6910, 1984. Belton, M.J.S., K.P. Klaasen, M.C. Clary, J.L. Anderson, C.D. Anger, M.H. Carr, C.R. Chapman, M.E. Davies, R. Greeley, D. Anderson, L.K. Bolef, T.E Townsend, R. Greenberg, J.W. Head, G. Neukum, C.B. Pileher, J. Veverka, P.J. Gierasch, F.P. Fanale, A.P. Ingersoll, H. Masursky, D. Morrison, and J.B. Pollack, The Galileo solidstate imaging experiment, Space Sci. Rev., 60, 413-455, 1992a. Belton, M. J. S., J.W. Head, C.M. Pieters, R. Greeley, A.S. McEwen, G. Neukum, K.P. Klaasen, C.D. Anger, M.H. Carr, C.R. Chapman, M.E. Davies, F.P. Fanale, P.J. Gierasch, R. Greenberg, A.P. Ingersoll, T. Johnson, B. Paczkowski, C.B. Pileher, and J. Veverka, Lunar impact basins and crustal heterogeneity: New western limb and farside data from Galileo, Science, 255, 570-576, 1992b. Bielefeld, M.J., R.C. Reedy, A.E. Metzger, J.I. Trombka, and J.R. Arnold, Surface chemistry of selected lunar regions, Proc. Lunar Planet. Sci. Conf., 7th, (3), 2661-2676, 1976. Bills, B.G., and A.J. Ferrari, Lunar crustal thickness, Proc. Lunar Planet. Sci. Conf., 7, frontispiece, 1976. Bratt, S.R., S.C. Solomon, and J.W. Head, The evolution of impact basins: Cooling, subsidence and thermal stress, J. Geophys. Res., 90, 12,415-12,433, 1985. Chao, E.C.T., Geologic implications of the Apollo 14 Fra Mauro breccias and comparison with ejecta from the Ries crater, Germany, U.S. Geol. Surv. J. Res., 1, 1-18, 1973. Clark, P.E., and B.R. Hawke, The lunar farside: The nature of highlands east of Mare Smythii, Earth Moon Planets, 53, 93-107, 1991. Davis, P.A., Jr., Iron and titanium distribution on the moon from orbital gamma ray spectrometry with implications for crustal evolutionary models, J. Geophys. Res., 85, 3209-3224, 1980. Davis, P.A., and P.D. Spudis, Petrologic province maps of the lunar highlands derived from orbital geochemical data, Proc. Lunar Planet. Sci. Conf. 16th, Part 1, J. Geophys. Res., 90, suppl., D61- D74, 1985. Davis, P.A., and P.D. Spudis, Global petrologic variations on the Moon: A ternary-diagram approach, Proc. Lunar Planet. Sci. Conf. 17th, Part 2, J. Geophys. Res., 92, suppl., E387-E395, 1987. Eggleton, R.E., and G.G. Schaber, Cayley formation interpreted as basin ejecta, part B of Photogeology, Apollo 16 Preliminary Science Report, NASA Spec. Publ., SP-315, 29-7 to 29-16, 1972. Ferrari, A.J., Lunar gravity: A harmonic analysis, J. Geophys. Res., 82, 3065-3084, 1977. Gaddis, L.R., and J.W. Head, Distribution and volumes of lava ponds in the Orientale region of the Moon, Lunar Planet. Sci. Conf. 12th, 321-323, 1981. Gillespie, A. R., M.0. Smith, J.B. Adams, S.C. Willis, A.F. Fischer III, and D.E. Sabol, Interpretation of residual images: Spectral mixture analysis of AVIRIS images, Owens Valley, California, Proc. of the Second Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) Workshop, edited by R.O. Green, JPL Publ., 90-54, 243-270, 1990.
17,180 HEAD ET AL.: LUNAR IMPACT BASINS Greeley, R., S.D. Kadel, D.A. Williams, L.R. Gaddis, J.W. Head, A.S. McEwen, S.L. Murchie, E. Nagel, G. Neukum, C.M. Pieters, J.M. Sunshine, R. Wagner, and M.J.S. Belton, Galileo imaging observations of lunar maria and related deposits, J. Geophys. Res., this issue. Grieve, R.A.F., Terrestrial impact: The record in the rocks, Meteoritics, 26, 175-194, 1991. Grieve, R.A.F., P.B. Robertson, and M.R. Dence, Constraints on the formation of ring impact structures, based on terrestrial data, in Multi-Ring Basins, edited by P.H. Schultz and R.B. Merrill, pp. 37-57, Pergamon, New York, 1981. Grieve, R.A.F., D. Stoffier, and A. Deutsch, The Sudbury structure: Controversial or misunderstood?, J. Geophys. Res., 96, 22,753-22,764, 1991. Hartmann, W.K., and G.P. Kuiper, Concentric structures surrounding lunar basins, Lunar Planet. Lab. Commun., 1, pp. 51-66, Univ. of Ariz., Tucson, 1962. Hartmann, W.K., and C.A. Wood, Moon: Origin and evolution of multi-ring basins, Moon, 3, 3-78, 1971. Hawke, B., and J. Bell, Remote sensing studies of lunar dark-halo impact craters: Preliminary results and implications for early volcanism, Proc. Lunar Planet. Sci. Conf, 12th, 665-678, 1981. Hawke, B.R., and J.W. Head, Lunar KREEP volcanism: Geologic evidence for history and mode of emplacement, Proc. Lunar Planet. Sci. Conf, 9th, (3), 3285-3309, 1978. Hawke, B., P. Lucey, G. Taylor, J. Bell, C. Peterson, D. Blewett, K. Horton, G. Smith, and P. Spudis, Remote sensing studies of the Orientale region of the moon: A pre-galileo view, Geophys. Res. Lett., 18, 2141-2144, 1991. Hawke, B.R., P.G. Lucey, G.J. Taylor, and C.A. Peterson, The distribution and modes of occurrence of lunar anorthosite, Proc. Lunar Planet. Sci. Conf, 23rd, 505-506, 1992. Head, J., Orientale multi-ring basin interior and implications for the petrogenesis of lunar highland samples, Moon, 11, 324-356, 1974a. Head, J. W., Stratigraphy of the Descartes region (Apollo 16): Implications for the origin of samples, Moon, 11, 77-99, 1974b. Head, J.W., Processes of lunar crater degradation: Changes in style with geologic time, Moon, 12, 299-329, 1975. Head, J.W., Origin of outer rings in lunar multi-ringed basins: Evidence from morphology and ring spacing, in Impact and Explosion Cratering, edited by D. Roddy, R. Pepin, and R. Merrill, pp. 563-573, Pergamon, New York, 1977. Head, J.W., Lava flooding of ancient planetary crusts: Geometry, thickness, and volumes of flooded lunar impact basins, Moon Planets, 26, 61-88, 1982. Head, J. W., and L. Wilson, Lunar mare volcanism: Stratigraphy, eruption conditions, and the evolution of secondary crusts, Geochim. Cosmochim. Acta, 56, 2155-2175, 1992. Helfenstein, P., and J. Veverka, Photometric properties of lunar terrains derived from Hapke's equation, Icarus, 72, 342-357, 1987. Hodges, C., and D. Wilhelms, Formation of lunar basin rings, Icarus, 34, 294-323, 1978. Howard, K.A., D.E. Wilhelms, and D.H. Scott, Lunar basins formation and highland stratigraphy, Rev. Geophys., 12, 309-327, 1974. Hurtrez, S., C. Sotin, S. Erard, J.-P. Bibring, J. Mustard, and C. Pieters, Evidence of mixing processes on Mars by studying ISM data, Proc. Lunar Planet. Sci. Conf, 22nd, 599-600, 1991. Kaula, W.M., G. Schubert, R.E. Lingenfelter, W.L. Sjogren, and W.R. Wollenhaupt, Lunar topography from Apollo 15 and 16 laser altimetry, Proc. Lunar Planet. Sci. Conf, 4th, (3), 2811-2819, 1973. Kislyuk, V.S., Profile of the far side of the Moon using Zond-8 data, Astrometriya Astrofiz., 31-36, 1975. Leikin, G.A., and A.N. $anovich, Origin of the southern basin on the far side of the Moon, Astron. Vestnik, 19, 113-119, 1985. McCauley, J., Orientale and Caloris, Phys. Earth Planet. Inter., 15, 220-250, 1977. McEwen, A.$., L.R. Gaddis, G. Neukum, H. Hoffmann, C.M. Pieters, and J.W. Head, Galileo observations of post-imbrium lunar craters during the first Earth-Moon flyby, J. Geophys. Res., this issue. Melosh, H.J., Impact Cratering: A Geologic Process, 254 pp., Oxford University Press, New York, 1989. Metzger, A.E., J.I. Trombka, L.E. Peterson, J.C. Reed, and J.R. Arnold, Lunar surface radioactivity: Preliminary results of the Apollo 15 and Apollo 16 gamma-ray spectrometer experiments, Science, 179, 800-803, 1973a. Metzger, A.E., J.I. Trombka, L.E. Peterson, J.C. Reed, and J.R. Arnold, Net lunar radioactivity, Geochim. Cosmochim. Acta, suppl. 4, 1, frontispiece, pl. 2, 1973b. Metzger, A.E., J.I. Trombka, R,C. Reedy, and J.R. Arnold, Element concentrations from lunar orbital gamma-ray measurements, Geochim. Cosmochim. Acta, suppl. 5, 2, 1067-1078, frontispiece, pl. 2, 1974. Metzger, A.E., E.L. Haines, R.E. Parker, and R.G. Radocinski, Thorium concentrations in the lunar surface, I, Regional values and crustal content, Proc. Lunar Planet. Sci. Conf., 8th, (1), 949-999, 1977. Moore, H., C. Hodges, and D. Scott, Lunar multiringed basinsillustrated by Oftentale and associated features, Proc. Lunar Sci. Conf., loth, 71-100, 1974. Murchie, $., J. Head, A. McEwen, J. Mustard, C. Pieters, and M. Belton, Spectral properties of Orientale basin materials from Galileo images, Proc. Lunar Planet. Sci. Conf., 23rd, 947-948, 1992. Mustard, J.F., and C.M. Pieters, Photometric phase functions of common geologic minerals and applications to quantitative analysis of mineral mixture reflectance spectra, J. Geophys. Res., 94, 13,619-13,634, 1989. Mustard, J., J. Head, S. Murchie, C. Pieters, M. Belton, and A. McEwen, Schickard cryptomare: Interaction between Orientale ejecta and pre-basin mare from spectral mixture analysis of Galileo $$I data, Proc. Lunar Planet. Sci. Conf., 23rd, 957-958, 1992. Oberbeck, V.R., The role of ballistic erosion and sedimentation in lunar stratigraphy, Rev. Geophys., 13, 337-362, 1975. Oberbeck, V.R., F. H/Srz, R.H. Morrison, W.L. Quaide, and D.E. Gault, On the origin of the lunar smooth-plains, Moon, 12, 19-54, 1975. Pieters, C.M., Copernicus crater central peak: Lunar mountain of unique composition, Science, 215, 59-61, 1982. Pieters, C. M., Composition of the lunar highland crust from nearinfrared spectroscopy, Rev. Geophys., 24, 557-578, 1986. Pieters, C. M., J.B. Adams, P.J. Mouginis-Mark, $.H. Zisk, M.O. Smith, J.W. Head, and T.B. McCord, The nature of crater rays: The Copernicus example, J. Geophys., Res., 90, 12,393-12,413, 1985. Pieters, C. M., M. Belton, E. Fischer, R. Greeley, R. Jaumann, J.W. Head, H. Hoffmann, A. McEwen, $. Murchie, G. Neukum, and J. Sunshine, Compositional implications of $$I multispectral images of the unexplored Lunar Farside, Proc. Lunar Planet. Sci. Conf., 23rd, 1069-1070, 1992. Pieters, C.M., J.W. Head, J.S. Sunshine, E. Fischer, S. Murchie, M. Belton, A. McEwen, L. Gaddis, R. Greeley, G. Neukum, R. Jaumann, and H. Hoffman, Crustal diversity of the Moon: Compositional analyses of Galileo solid state imaging data, J. Geophys. Res., this issue. Robinson, M.$., B.R. Hawke, P.G. Lucey, and G.A. Smith, Mariner 10 multispectral images of the eastern limb and farside of the Moon, J. Geophys. Res., 97, 18,265-18,274, 1992. Rodionov, B.N., et al., New information on lunar relief from Zond-6 photographs, Kosm. Issled., 9, 450-458, 1971. Rodionov, B.N., A.A. Nefed'ev, M.I. $hpekin, et al., Study of the relief on the far side of the Moon using Zond-8 photographs, Kosm. Issled., 14, 624-629, 1976. Ryder, G., and J.A. Wood, Serenitatis and Imbrium impact melts: Implications for large-scale layering in the lunar crust, Proc. Lunar Planet. Sci. Conf., 8th, 655-668, 1977. Schultz, P.H., Moon Morphology, University of Texas Press, Austin, 1972. Schultz, P.H., and D.E. Gault, Clustered impacts: Experiments and implications, J. Geophys. Res., 90, 3701-3732, 1985. Schultz, P.H, and P.A. Spudis, Evidence for ancient mare volcanism, Proc. Lunar Planet. Sci. Conf, loth, 2899-2918, 1979. Scott, D.H., J.M. Diaz, and J.A. Watkins, Lunar farside tectonics and volcanism, Proc. Lunar Planet. Sci. Conf, 8th, (1), 1119-1130, 1977. Shevchenko, V.V., Modern Study of the Moon, 284 pp., Moscow Science, 1980. Shevchenko, V.V., and V.I. Chikmachev, Western hemisphere of the Moon: "Zond" and "Galileo" views, Vernadsky/Brown Microsyrup., 16, 81-82, 1992. Shpekin, M.I., Relief of Mare Orientale on the Moon on the basis of "Zond-6, -8" data, Candidate thesis, 142 pp., Kazan Univ., Kazan, Russian Federation, 1983.
HEAD ET AL.: LUNAR IMPACt BASINS 17,181 Singer, R.B., and T.B. McCord, Mars: Large scale mixing of bright and dark surface materials and implications for analysis of spectral Wilhelms, D.E., D.E. Stuart-Alexander, and K.A. Howard, Preliminary interpretations of lunar geology, in Initial Photographic Analysis, reflectance, Proc. Lunar Planet. Sci. Conf., loth, 1835-1848, chap. 2, Analysis of Apollo 8 Photography and Visual 1979. Observations, NASA Spec. Rep. SP-201, 16-21, 1969. Soderblom, L.A., and L.A. Lebofsky, Technique for rapid determination of relative ages of lunar areas from orbital photography, J. Geophys. Res., 77, 279-296, 1972. Solomon, S.C., R.P. Comer, and J.W. Head, The evolution of impact basins: Viscous relaxation of topographic relief, J. Geophys. Res., Wilhelms, D., C. Hodges, and R. Pike, Nested crater model of lunar ringed basins, in Impact and Explosion Cratering, edited by D. Roddy, R. Pepin, and R. Merrill, pp. 539-562, Pergamon, New York, 1977. Wilhelms, D.E., K.A. Howard, and H.G. Wilshire, Geologic map of the 87, 3975-3992, 1982. south side of the Moon, U.S. Geol. Surv. Map, 1-1162, 1979. Spudis, P.D., and P.A. Davis, A chemical and petrological model of the lunar crust and implications for lunar crustal origin, Proc. Lunar Wollenhaupt, W.R., and W.L. Sjogren, Comments on the figure of the Moon based on preliminary results from laser altimetry, Moon, 4, Planet. Sci. Conf. 17th, Part 1, J. Geophys. Res., 91, suppl., E84-337-347, 1972a. E90, 1986. Wollenhaupt, W.R., and W.L. Sjogren, Apollo 16 laser altimeter, Spudis, P.D, B. Hawke, and P. Lucey, Composition of Orientale basin deposits and implications for the lunar basin-forming process, Proc. Lunar Planet. Sci. Conf. 15th, J. Geophys. Res., 89, suppl., Apollo 16 Preliminary Science Report, NASA Spec. Publ., SP-315, 30-1 to 30-5, 1972b. Wollenhaupt, W.R., W.L. Sjogren, R.E. Lingenfelter, G. Schubert, and C197-210, 1984. W.M. Kaula, Apollo 17 laser altimeter, Apollo 17 Preliminary Spudis, P.D., B.R. Hawke, and P.G. Lucey, Materials and formation of the Imbrium basin, Proc. Lunar Planet. Sci. Conf., 18th, 155-168, 1988. Stuart-Alexander, D.E., Geologic map of the central farside of the Moon, U.S. Geol. Surv., Misc Inv. Ser. Map, 1-1047, 1978. Stuart-Alexander, D.E., and K.A. Howard, Lunar maria and circular basins-a review, Icarus, 12, 440-456, 1970. Science Report, NASA Spec. Publ., SP-330, 33-41 to 33-44, 1973. Wood, C.A., and A.W. Gifford, Evidence for the lunar big backside basin, Multi-Ring Basins: Formation and Evolution Conference, Rep. 414, pp. 121-123, Lunar and Planet. Inst., Houston, Tex., 1980a. Wood, C.A., and A.W. Gifford, Crater distributions and the evolution of the lunar farside highlands, NASA Tech. Memo., TM-81776, Trombka, J.I., J.R Arnold, R.C. Reedy, and L.E. Peterson, Some correlations between measurements by the Apollo gamma-ray 111-118, 1980b. spectrometer and other observations, Geochim. Cosmochim. Acta, Suppl. 4, 3, 2847-2853, 1973. Warren, P.H., Lunar anorthosites and the magma-ocean plagioclaseflotation hypothesis: Importance of FeO enrichment in the parent magma, Am. Mineral., 75, 46-58, 1990. Whitaker, E., The lunar Procellarum basin, Multi-Ring Basins, Proc. Lunar Planet. Sci. Conf., 12th, Part. A, 105-111, 1981. Whitford-Stark, J.L., Charting the Southern Seas: The evolution of the lunar Mare Australe, Proc. Lunar Planet. Sci. Conf., loth, 2975-2994, 1979. M. J. S. Belton, National Optical Astronomy Observatory, 950 N. Cherry Ave., Tucson, AZ 85726. R. Greeley, Department of Geology, Arizona State University, Tempe, AZ 85287. J. W. Head, S. Murchie, J. F. Mustard, and C. M. Pieters, Department of Geological Sciences, Brown University, Providence, RI 02912. A. McEwen, U.S. Geological Survey, 2255 N. Gemini Road, Flagstaff, AZ 86001. E. Nagel and G. Neukum, DLR, Institute for Planetary Exploration, 12484 Berlin and 82230 Oberpfaffenhofen, Germany. Wilhelms, D.E., The geology of the moon, U.S. Geol. Surv. Prof. Pap., 1348, 1987. Wilhelms, D.E., and F. E1-Baz, Geologic map of the east side of the Moon, scale 1:5,000,000, U.S. Geol. Surv. Map, 1-948, 1977. (Received September 10, 1992; revised May 12, 1993; accepted May 13 1993)